rust/lancedb
1
0
Fork 0
mirror of https://github.com/lancedb/lancedb.git synced 2026-10-21 19:22:27 +00:00
Code Issues Packages Projects Releases Wiki Activity

refactor!: rename MemWAL APIs to RTE (realtime engine) (#4598)

Browse Source 
## What

Renames the user-facing MemWAL surface to **RTE** (realtime engine)
across Rust, Python, TypeScript, and Java, including the REST wire
format. This follows the LSM → MemWAL rename in #4075.

| Before | After |
|---|---|
| `MemWalWriteSpec` / `MemWalStats` | `RTEWriteSpec` / `RTEStats` |
| `enable_memwal` / `disable_memwal` / `get_memwal_spec` | `enable_rte`
/ `disable_rte` / `get_rte_spec` |
| `flush_memwal` / `compact_memwal` / `checkpoint_memwal` | `flush_rte`
/ `compact_rte` / `checkpoint_rte` |
| `get_memwal_stats` / `close_memwal_writers` | `get_rte_stats` /
`close_rte_writers` |
| `use_memwal` (query, merge insert) | `use_rte` |
| TS/Java `enableMemWal`, `useMemWal`, ... | `enableRTE`, `useRTE`, ...
(napi `js_name` set explicitly) |
| Java `LanceDbTableMemWal` | `LanceDbTableRTE` |
| `table/memwal_stats.rs` | `table/rte_stats.rs` |

Docs and comments now say "realtime engine (RTE)". Text about Lance's
underlying mechanism keeps "MemWAL" ("Lance's MemWAL", "MemWAL index").

**Left alone:** Lance-owned names (`MemWalIndexDetails`,
`DatasetMemWalExt`, `InitializeMemWal*`, `mem_wal` paths), internal
`Lsm*` names and modules, and the `does not support with_row_id` refusal
text that the hybrid fallback matches on.

## Compatibility

Breaking, for an experimental API. The wire format changes, so this must
roll out together with the matching sophon change (lancedb/sophon#8870):

- Routes: `/enable_memwal/` → `/enable_rte/`, `/disable_memwal/` →
`/disable_rte/`, `/get_memwal_spec/` → `/get_rte_spec/`,
`/get_memwal_stats/` → `/get_rte_stats/`, `/flush_memwal/` →
`/flush_rte/`, `/compact_memwal/` → `/compact_rte/`
- Response keys: `memwal_spec` → `rte_spec`, `memwal_stats` →
`rte_stats`
- Request field: `use_memwal` → `use_rte` on query and merge-insert
bodies

The SQL keyword becomes `ALTER TABLE ... SET RTE` server-side. The
guides under `docs/web` will be updated separately.

## Validation

- Rust: `cargo clippy --features remote --tests --examples` clean;
`cargo test --features remote -p lancedb --lib`: 1533 passed
- Java: `mvn test`: 37 passed (Spotless applied with JDK 11)
- Python: `ruff format` / `ruff check` clean; targeted tests
(`test_rte_spec`, `test_merge_insert_lsm`, `test_remote_db`,
`test_wal_hybrid`, `test_query`, `test_permutation`, `test_table`): 708
passed
- TypeScript: `pnpm build`; `table`, `query` tests: 606 passed; `biome
check` clean; `docs/src/js` regenerated with `pnpm run docs`

🤖 Generated with [Claude Code](https://claude.com/claude-code)

---------

Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
This commit is contained in:
Dan Rammer
2026-10-20 11:58:57 -05:00
committed by GitHub
co-authored by Claude Opus 5.5
parent 1772f4ef83
commit 92825f6f13
73 changed files with 1656 additions and 1600 deletions
Download Patch File Download Diff File Expand all files Collapse all files
docs/src/js/classes/AutoQuery.md
+10 -10
View File
@@ -425,16 +425,16 @@ Empty results retain the query's output schema.
***
### useMemWal()
### useRTE()
```ts
useMemWal(enable): this
useRTE(enable): this
```
Control MemWAL read routing for this query.
Control RTE read routing for this query.
By default (unset), when the table carries a MemWAL write spec (see
[Table#enableMemWal](Table.md#enablememwal)), reads are routed through the LSM scanner so
By default (unset), when the table carries an RTE write spec (see
[Table#enableRTE](Table.md#enablerte)), reads are routed through the LSM scanner so
they also return data written via the `mergeInsert` LSM path that has not yet
been compacted into the base table (the active/frozen in-memory memtables and
the SSTables), deduplicated by primary key; a table without a spec
@@ -444,12 +444,12 @@ reads the base table.
* **enable**: `boolean`
`true` forces the LSM scanner and errors if the table has no
MemWAL write spec. `false` bypasses the MemWAL and reads the base table only,
RTE write spec. `false` bypasses the RTE and reads the base table only,
even when a spec is present.
Note: the LSM scanner does not support every query shape (e.g. reranking,
hybrid search, `orderBy`). On a MemWAL table those shapes error unless
`useMemWal(false)` is set, because a base-only read would silently exclude
un-compacted MemWAL data.
hybrid search, `orderBy`). On an RTE table those shapes error unless
`useRTE(false)` is set, because a base-only read would silently exclude
un-compacted RTE data.
#### Returns
@@ -457,7 +457,7 @@ reads the base table.
#### Inherited from
`StandardQueryBase.useMemWal`
`StandardQueryBase.useRTE`
***
docs/src/js/classes/MergeInsertBuilder.md
+8 -8
View File
@@ -76,23 +76,23 @@ the query optimizer chooses a suboptimal path.
***
### useMemWal()
### useRTE()
```ts
useMemWal(enable): MergeInsertBuilder
useRTE(enable): MergeInsertBuilder
```
Control MemWAL routing for this merge.
Control RTE routing for this merge.
By default (unset), a `mergeInsert` on a table with a MemWAL write spec is
By default (unset), a `mergeInsert` on a table with an RTE write spec is
routed through Lance's MemWAL shard writer, and a table without one uses the
standard path.
#### Parameters
* **enable**: `boolean`
`true` forces MemWAL routing and errors if the table has no
MemWAL write spec. `false` forces the standard write path even when a spec is set.
`true` forces RTE routing and errors if the table has no
RTE write spec. `false` forces the standard write path even when a spec is set.
#### Returns
@@ -108,12 +108,12 @@ validateSingleShard(validateSingleShard): MergeInsertBuilder
Controls how an LSM merge checks that its input targets a single shard.
When a table has a MemWAL write spec, every row in a `mergeInsert` call must
When a table has an RTE write spec, every row in a `mergeInsert` call must
route to the same shard. When `true` (the default), every row is inspected
to verify this. When `false`, only the first row is inspected and the
shard it routes to is used for the whole input — a faster path for callers
that have already pre-sharded their input. Has no effect on tables without
a MemWAL write spec.
an RTE write spec.
#### Parameters
docs/src/js/classes/Query.md
+10 -10
View File
@@ -499,16 +499,16 @@ Empty results retain the query's output schema.
***
### useMemWal()
### useRTE()
```ts
useMemWal(enable): this
useRTE(enable): this
```
Control MemWAL read routing for this query.
Control RTE read routing for this query.
By default (unset), when the table carries a MemWAL write spec (see
[Table#enableMemWal](Table.md#enablememwal)), reads are routed through the LSM scanner so
By default (unset), when the table carries an RTE write spec (see
[Table#enableRTE](Table.md#enablerte)), reads are routed through the LSM scanner so
they also return data written via the `mergeInsert` LSM path that has not yet
been compacted into the base table (the active/frozen in-memory memtables and
the SSTables), deduplicated by primary key; a table without a spec
@@ -518,12 +518,12 @@ reads the base table.
* **enable**: `boolean`
`true` forces the LSM scanner and errors if the table has no
MemWAL write spec. `false` bypasses the MemWAL and reads the base table only,
RTE write spec. `false` bypasses the RTE and reads the base table only,
even when a spec is present.
Note: the LSM scanner does not support every query shape (e.g. reranking,
hybrid search, `orderBy`). On a MemWAL table those shapes error unless
`useMemWal(false)` is set, because a base-only read would silently exclude
un-compacted MemWAL data.
hybrid search, `orderBy`). On an RTE table those shapes error unless
`useRTE(false)` is set, because a base-only read would silently exclude
un-compacted RTE data.
#### Returns
@@ -531,7 +531,7 @@ reads the base table.
#### Inherited from
`StandardQueryBase.useMemWal`
`StandardQueryBase.useRTE`
***
docs/src/js/classes/Table.md
+37 -37
View File
@@ -228,13 +228,13 @@ version of the table.
***
### checkpointMemWal()
### checkpointRTE()
```ts
abstract checkpointMemWal(): Promise<void>
abstract checkpointRTE(): Promise<void>
```
Converge this table's MemWAL write path into its base table.
Converge this table's RTE write path into its base table.
Freezes once, then triggers compaction and polls until the SSTables that existed
at the start is gone. The target set is fixed at the start, so
@@ -254,9 +254,9 @@ is merging. The caller owns the deadline.
#### Example
```ts
const before = await table.getMemWalStats();
await table.checkpointMemWal();
const after = await table.getMemWalStats();
const before = await table.getRTEStats();
await table.checkpointRTE();
const after = await table.getRTEStats();
```
***
@@ -279,15 +279,15 @@ Any attempt to use the table after it is closed will result in an error.
***
### closeMemWalWriters()
### closeRTEWriters()
```ts
abstract closeMemWalWriters(): Promise<void>
abstract closeRTEWriters(): Promise<void>
```
Drain and close any cached MemWAL shard writers held for this table.
Drain and close any cached RTE shard writers held for this table.
When an [MemWalWriteSpec](../interfaces/MemWalWriteSpec.md) is installed, `mergeInsert` opens MemWAL
When an [RTEWriteSpec](../interfaces/RTEWriteSpec.md) is installed, `mergeInsert` opens RTE
shard writers and caches them for reuse across calls. This closes them,
flushing pending data; writers reopen lazily on the next `mergeInsert`.
It is a no-op when no writers are cached.
@@ -298,17 +298,17 @@ It is a no-op when no writers are cached.
***
### compactMemWal()
### compactRTE()
```ts
abstract compactMemWal(): Promise<void>
abstract compactRTE(): Promise<void>
```
Trigger a background SSTable compaction pass per table shard.
Returns once the passes are *dispatched*, not once they finish — watch
[Table#getMemWalStats](Table.md#getmemwalstats) for progress, or use
[Table#checkpointMemWal](Table.md#checkpointmemwal) to wait for convergence.
[Table#getRTEStats](Table.md#getrtestats) for progress, or use
[Table#checkpointRTE](Table.md#checkpointrte) to wait for convergence.
#### Returns
@@ -449,13 +449,13 @@ containing the new version number of the table
***
### disableMemWal()
### disableRTE()
```ts
abstract disableMemWal(): Promise<void>
abstract disableRTE(): Promise<void>
```
Remove the [MemWalWriteSpec](../interfaces/MemWalWriteSpec.md) from this table, reverting to the standard
Remove the [RTEWriteSpec](../interfaces/RTEWriteSpec.md) from this table, reverting to the standard
`mergeInsert` write path.
Errors if no spec is currently set.
@@ -531,16 +531,16 @@ Drop an index from the table.
***
### enableMemWal()
### enableRTE()
```ts
abstract enableMemWal(spec): Promise<void>
abstract enableRTE(spec): Promise<void>
```
Install an [MemWalWriteSpec](../interfaces/MemWalWriteSpec.md) on this table, selecting Lance's MemWAL
Install an [RTEWriteSpec](../interfaces/RTEWriteSpec.md) on this table, selecting Lance's MemWAL
LSM-style write path for future `mergeInsert` calls.
`MemWalWriteSpec` chooses one of three sharding strategies via `specType`:
`RTEWriteSpec` chooses one of three sharding strategies via `specType`:
- `"bucket"` — hash-bucket writes by the single-column unenforced primary
key (`column` and `numBuckets` required).
@@ -552,13 +552,13 @@ All variants require the table to have an unenforced primary key
requires it to be the single column being bucketed.
Omitting `maintainedIndexes` maintains every index the table has,
including ones created later, and skips a kind the MemWAL cannot maintain.
including ones created later, and skips a kind the RTE cannot maintain.
Naming them pins an exact set, and a still-building index is rejected
rather than quietly omitted.
#### Parameters
* **spec**: [`MemWalWriteSpec`](../interfaces/MemWalWriteSpec.md)
* **spec**: [`RTEWriteSpec`](../interfaces/RTEWriteSpec.md)
The sharding spec to install.
#### Returns
@@ -569,7 +569,7 @@ rather than quietly omitted.
```ts
await table.setUnenforcedPrimaryKey("id");
await table.enableMemWal({
await table.enableRTE({
specType: "bucket",
column: "id",
numBuckets: 16,
@@ -647,10 +647,10 @@ memory; use [Table.fetchBlobFiles](Table.md#fetchblobfiles) for large values.
***
### flushMemWal()
### flushRTE()
```ts
abstract flushMemWal(): Promise<void>
abstract flushRTE(): Promise<void>
```
Freeze every table shard's active memtable into a new SSTable.
@@ -699,38 +699,38 @@ const { records, truncated } = await table.functionErrors({ column: "embedding"
***
### getMemWalSpec()
### getRTESpec()
```ts
abstract getMemWalSpec(): Promise<undefined | MemWalWriteSpec>
abstract getRTESpec(): Promise<undefined | RTEWriteSpec>
```
Read the [MemWalWriteSpec](../interfaces/MemWalWriteSpec.md) currently installed on this table.
Read the [RTEWriteSpec](../interfaces/RTEWriteSpec.md) currently installed on this table.
Resolves to `undefined` when the MemWAL write path is not enabled (no
spec has been set, or it was removed with [Table#disableMemWal](Table.md#disablememwal)).
Resolves to `undefined` when the RTE write path is not enabled (no
spec has been set, or it was removed with [Table#disableRTE](Table.md#disablerte)).
The spec is the one installed, including its maintained-index selection:
an absent `maintainedIndexes` for every index the table has, an empty
array for none.
#### Returns
`Promise`&lt;`undefined` \| [`MemWalWriteSpec`](../interfaces/MemWalWriteSpec.md)&gt;
`Promise`&lt;`undefined` \| [`RTEWriteSpec`](../interfaces/RTEWriteSpec.md)&gt;
***
### getMemWalStats()
### getRTEStats()
```ts
abstract getMemWalStats(includeSstableRows?): Promise<undefined | MemWalStats>
abstract getRTEStats(includeSstableRows?): Promise<undefined | RTEStats>
```
Read live per-table-shard MemWAL state.
Read live per-table-shard RTE state.
Answers "how far behind is my fresh tier", "which table shard is hot", and
"why is my fresh-tier vector search brute-force". Mutates no table state.
Resolves to `undefined` only when the MemWAL write path is not enabled.
Resolves to `undefined` only when the RTE write path is not enabled.
#### Parameters
@@ -740,7 +740,7 @@ Resolves to `undefined` only when the MemWAL write path is not enabled.
#### Returns
`Promise`&lt;`undefined` \| [`MemWalStats`](../interfaces/MemWalStats.md)&gt;
`Promise`&lt;`undefined` \| [`RTEStats`](../interfaces/RTEStats.md)&gt;
***
docs/src/js/classes/TakeQuery.md
+5 -5
View File
@@ -275,17 +275,17 @@ Empty results retain the query's output schema.
***
### useMemWal()
### useRTE()
```ts
useMemWal(enable): this
useRTE(enable): this
```
Control MemWAL read routing for this take query.
Control RTE read routing for this take query.
`false` bypasses the MemWAL and reads the base table only — the escape hatch,
`false` bypasses the RTE and reads the base table only — the escape hatch,
since take-by-row-id/offset is not supported on the LSM scanner and, on a
MemWAL table, auto-routes to it and errors otherwise.
RTE table, auto-routes to it and errors otherwise.
#### Parameters
docs/src/js/classes/VectorQuery.md
+11 -11
View File
@@ -659,7 +659,7 @@ minimum; one searches all partitions within the caller maximum. Later
probing may continue when too few candidates are found.
Any explicit effort, including 0.5, overrides nprobes in either call order.
Explicit minimum and maximum bounds still apply. This is not a recall
guarantee and does not change refinement or HNSW ef. The MemWAL LSM
guarantee and does not change refinement or HNSW ef. The RTE LSM
scanner does not yet support this option.
#### Parameters
@@ -781,16 +781,16 @@ Empty results retain the query's output schema.
***
### useMemWal()
### useRTE()
```ts
useMemWal(enable): this
useRTE(enable): this
```
Control MemWAL read routing for this query.
Control RTE read routing for this query.
By default (unset), when the table carries a MemWAL write spec (see
[Table#enableMemWal](Table.md#enablememwal)), reads are routed through the LSM scanner so
By default (unset), when the table carries an RTE write spec (see
[Table#enableRTE](Table.md#enablerte)), reads are routed through the LSM scanner so
they also return data written via the `mergeInsert` LSM path that has not yet
been compacted into the base table (the active/frozen in-memory memtables and
the SSTables), deduplicated by primary key; a table without a spec
@@ -800,12 +800,12 @@ reads the base table.
* **enable**: `boolean`
`true` forces the LSM scanner and errors if the table has no
MemWAL write spec. `false` bypasses the MemWAL and reads the base table only,
RTE write spec. `false` bypasses the RTE and reads the base table only,
even when a spec is present.
Note: the LSM scanner does not support every query shape (e.g. reranking,
hybrid search, `orderBy`). On a MemWAL table those shapes error unless
`useMemWal(false)` is set, because a base-only read would silently exclude
un-compacted MemWAL data.
hybrid search, `orderBy`). On an RTE table those shapes error unless
`useRTE(false)` is set, because a base-only read would silently exclude
un-compacted RTE data.
#### Returns
@@ -813,7 +813,7 @@ reads the base table.
#### Inherited from
`StandardQueryBase.useMemWal`
`StandardQueryBase.useRTE`
***
docs/src/js/globals.md
+2 -2
View File
@@ -114,8 +114,6 @@
- [ListTablesResponse](interfaces/ListTablesResponse.md)
- [ListingOptions](interfaces/ListingOptions.md)
- [MaterializedViewDefinition](interfaces/MaterializedViewDefinition.md)
- [MemWalStats](interfaces/MemWalStats.md)
- [MemWalWriteSpec](interfaces/MemWalWriteSpec.md)
- [MemtableStats](interfaces/MemtableStats.md)
- [MergeResult](interfaces/MergeResult.md)
- [NativeOAuthConfig](interfaces/NativeOAuthConfig.md)
@@ -124,6 +122,8 @@
- [OptimizeOptions](interfaces/OptimizeOptions.md)
- [OptimizeStats](interfaces/OptimizeStats.md)
- [QueryExecutionOptions](interfaces/QueryExecutionOptions.md)
- [RTEStats](interfaces/RTEStats.md)
- [RTEWriteSpec](interfaces/RTEWriteSpec.md)
- [RefreshColumnResult](interfaces/RefreshColumnResult.md)
- [RefreshMaterializedViewResult](interfaces/RefreshMaterializedViewResult.md)
- [RemovalStats](interfaces/RemovalStats.md)
docs/src/js/interfaces/MemWalStats.md → docs/src/js/interfaces/RTEStats.md
+3 -3
View File
@@ -2,11 +2,11 @@
***
[@lancedb/lancedb](../globals.md) / MemWalStats
[@lancedb/lancedb](../globals.md) / RTEStats
# Interface: MemWalStats
# Interface: RTEStats
Live per-table-shard MemWAL state, as returned by `Table#getMemWalStats`.
Live per-table-shard RTE state, as returned by `Table#getRTEStats`.
Nothing here is derived: sums and differences (total SSTable bytes, WAL lag) are
the caller's to compute.
docs/src/js/interfaces/MemWalWriteSpec.md → docs/src/js/interfaces/RTEWriteSpec.md
+5 -5
View File
@@ -2,12 +2,12 @@
***
[@lancedb/lancedb](../globals.md) / MemWalWriteSpec
[@lancedb/lancedb](../globals.md) / RTEWriteSpec
# Interface: MemWalWriteSpec
# Interface: RTEWriteSpec
Specification selecting Lance's MemWAL LSM-style write path for
`mergeInsert`.
Specification enabling the realtime engine (RTE), which routes
`mergeInsert` through Lance's MemWAL LSM-style write path.
`specType` is `"bucket"`, `"identity"`, or `"unsharded"`. For `"bucket"`,
`column` and `numBuckets` are required; for `"identity"`, `column` is
@@ -34,7 +34,7 @@ Bucket and identity variants: the sharding column.
optional maintainedIndexes: string[];
```
Indexes the MemWAL keeps up to date. Omit for every index the table has,
Indexes the RTE keeps up to date. Omit for every index the table has,
including ones created later. Pass `[]` for none.
***
docs/src/python/python.md
+1 -1
View File
@@ -297,7 +297,7 @@ listing a storage directory.
::: lancedb.table.Branches
::: lancedb.MemWalWriteSpec
::: lancedb.RTEWriteSpec
## Functions and Jobs
java/.project
+28
View File
@@ -0,0 +1,28 @@
<?xml version="1.0" encoding="UTF-8"?>
<projectDescription>
<name>lancedb-parent</name>
<comment></comment>
<projects>
</projects>
<buildSpec>
<buildCommand>
<name>org.eclipse.m2e.core.maven2Builder</name>
<arguments>
</arguments>
</buildCommand>
</buildSpec>
<natures>
<nature>org.eclipse.m2e.core.maven2Nature</nature>
</natures>
<filteredResources>
<filter>
<id>1786980408448</id>
<name></name>
<type>30</type>
<matcher>
<id>org.eclipse.core.resources.regexFilterMatcher</id>
<arguments>node_modules|\.git|__CREATED_BY_JAVA_LANGUAGE_SERVER__</arguments>
</matcher>
</filter>
</filteredResources>
</projectDescription>
java/.settings/org.eclipse.core.resources.prefs
+2
View File
@@ -0,0 +1,2 @@
eclipse.preferences.version=1
encoding/<project>=UTF-8
java/.settings/org.eclipse.m2e.core.prefs
+4
View File
@@ -0,0 +1,4 @@
activeProfiles=
eclipse.preferences.version=1
resolveWorkspaceProjects=true
version=1
java/README.md
+9 -9
View File
@@ -29,34 +29,34 @@ LanceNamespace namespaceClient = LanceDbNamespaceClientBuilder.newBuilder()
.build();
```
## MemWAL write path
## RTE write path
Most table operations reach LanceDB through the `LanceNamespace` above, which is
generated from the Lance Namespace specification. The MemWAL LSM routes are not part
generated from the Lance Namespace specification. The RTE LSM routes are not part
of that specification, so they are issued through a separate client:
```java
import com.lancedb.LanceDbRestClient;
import com.lancedb.LanceDbTableMemWal;
import com.lancedb.MemWalWriteSpec;
import com.lancedb.LanceDbTableRTE;
import com.lancedb.RTEWriteSpec;
LanceDbRestClient client = LanceDbNamespaceClientBuilder.newBuilder()
.apiKey("your_lancedb_cloud_api_key")
.database("your_database_name")
.buildRestClient();
LanceDbTableMemWal memWal = new LanceDbTableMemWal(client, "my_table");
LanceDbTableRTE rte = new LanceDbTableRTE(client, "my_table");
// Route future merge_insert upserts through the MemWAL, hash-bucketed by `id`.
memWal.enableMemWal(MemWalWriteSpec.bucket("id", 16));
// Route future merge_insert upserts through the RTE, hash-bucketed by `id`.
rte.enableRTE(RTEWriteSpec.bucket("id", 16));
// ... merge_insert traffic ...
// Converge the fresh tier into the base table.
memWal.checkpointMemWal();
rte.checkpointRTE();
// Inspect live per-table-shard state.
memWal.getMemWalStats().ifPresent(stats -> stats.tableShards().forEach(tableShard ->
rte.getRTEStats().ifPresent(stats -> stats.tableShards().forEach(tableShard ->
System.out.println(tableShard.shardId() + ": " + tableShard.sstables().size() + " SSTables")));
client.close();
java/lancedb-core/.classpath
+57
View File
@@ -0,0 +1,57 @@
<?xml version="1.0" encoding="UTF-8"?>
<classpath>
<classpathentry kind="src" output="target/classes" path="src/main/java">
<attributes>
<attribute name="optional" value="true"/>
<attribute name="maven.pomderived" value="true"/>
</attributes>
</classpathentry>
<classpathentry excluding="**" kind="src" output="target/classes" path="src/main/resources">
<attributes>
<attribute name="maven.pomderived" value="true"/>
<attribute name="optional" value="true"/>
</attributes>
</classpathentry>
<classpathentry kind="src" output="target/test-classes" path="src/test/java">
<attributes>
<attribute name="optional" value="true"/>
<attribute name="maven.pomderived" value="true"/>
<attribute name="test" value="true"/>
</attributes>
</classpathentry>
<classpathentry excluding="**" kind="src" output="target/test-classes" path="src/test/resources">
<attributes>
<attribute name="maven.pomderived" value="true"/>
<attribute name="test" value="true"/>
<attribute name="optional" value="true"/>
</attributes>
</classpathentry>
<classpathentry kind="con" path="org.eclipse.jdt.launching.JRE_CONTAINER/org.eclipse.jdt.internal.debug.ui.launcher.StandardVMType/JavaSE-11">
<attributes>
<attribute name="maven.pomderived" value="true"/>
</attributes>
</classpathentry>
<classpathentry kind="con" path="org.eclipse.m2e.MAVEN2_CLASSPATH_CONTAINER">
<attributes>
<attribute name="maven.pomderived" value="true"/>
</attributes>
</classpathentry>
<classpathentry kind="src" path="target/generated-sources/annotations">
<attributes>
<attribute name="optional" value="true"/>
<attribute name="maven.pomderived" value="true"/>
<attribute name="ignore_optional_problems" value="true"/>
<attribute name="m2e-apt" value="true"/>
</attributes>
</classpathentry>
<classpathentry kind="src" output="target/test-classes" path="target/generated-test-sources/test-annotations">
<attributes>
<attribute name="optional" value="true"/>
<attribute name="maven.pomderived" value="true"/>
<attribute name="ignore_optional_problems" value="true"/>
<attribute name="m2e-apt" value="true"/>
<attribute name="test" value="true"/>
</attributes>
</classpathentry>
<classpathentry kind="output" path="target/classes"/>
</classpath>
java/lancedb-core/.project
+34
View File
@@ -0,0 +1,34 @@
<?xml version="1.0" encoding="UTF-8"?>
<projectDescription>
<name>lancedb-core</name>
<comment></comment>
<projects>
</projects>
<buildSpec>
<buildCommand>
<name>org.eclipse.jdt.core.javabuilder</name>
<arguments>
</arguments>
</buildCommand>
<buildCommand>
<name>org.eclipse.m2e.core.maven2Builder</name>
<arguments>
</arguments>
</buildCommand>
</buildSpec>
<natures>
<nature>org.eclipse.jdt.core.javanature</nature>
<nature>org.eclipse.m2e.core.maven2Nature</nature>
</natures>
<filteredResources>
<filter>
<id>1786980408445</id>
<name></name>
<type>30</type>
<matcher>
<id>org.eclipse.core.resources.regexFilterMatcher</id>
<arguments>node_modules|\.git|__CREATED_BY_JAVA_LANGUAGE_SERVER__</arguments>
</matcher>
</filter>
</filteredResources>
</projectDescription>
java/lancedb-core/.settings/org.eclipse.core.resources.prefs
+5
View File
@@ -0,0 +1,5 @@
eclipse.preferences.version=1
encoding//src/main/java=UTF-8
encoding//src/test/java=UTF-8
encoding//src/test/resources=UTF-8
encoding/<project>=UTF-8
java/lancedb-core/.settings/org.eclipse.jdt.apt.core.prefs
+2
View File
@@ -0,0 +1,2 @@
eclipse.preferences.version=1
org.eclipse.jdt.apt.aptEnabled=false
java/lancedb-core/.settings/org.eclipse.jdt.core.prefs
+9
View File
@@ -0,0 +1,9 @@
eclipse.preferences.version=1
org.eclipse.jdt.core.compiler.codegen.targetPlatform=11
org.eclipse.jdt.core.compiler.compliance=11
org.eclipse.jdt.core.compiler.problem.enablePreviewFeatures=disabled
org.eclipse.jdt.core.compiler.problem.forbiddenReference=warning
org.eclipse.jdt.core.compiler.problem.reportPreviewFeatures=ignore
org.eclipse.jdt.core.compiler.processAnnotations=disabled
org.eclipse.jdt.core.compiler.release=disabled
org.eclipse.jdt.core.compiler.source=11
java/lancedb-core/.settings/org.eclipse.m2e.core.prefs
+4
View File
@@ -0,0 +1,4 @@
activeProfiles=
eclipse.preferences.version=1
resolveWorkspaceProjects=true
version=1
java/lancedb-core/src/main/java/com/lancedb/JsonFields.java
+2 -2
View File
@@ -20,11 +20,11 @@ import com.fasterxml.jackson.databind.JsonNode;
*
* <p>Every reader fails closed: a missing, null, or wrong-typed field throws rather than
* defaulting. That mirrors the serde decoding the Rust client applies to the same payloads in
* {@code rust/lancedb/src/table/memwal_stats.rs}, where a required field has no default and a
* {@code rust/lancedb/src/table/rte_stats.rs}, where a required field has no default and a
* malformed response is an error rather than a zero.
*
* <p>The alternative — Jackson's {@code path()}, which yields a missing node that reads as an empty
* array or a zero — is unsafe here because {@link LanceDbTableMemWal#checkpointMemWal()} decides
* array or a zero — is unsafe here because {@link LanceDbTableRTE#checkpointRTE()} decides
* convergence from these numbers. A defaulted {@code sstables} array is indistinguishable from a
* drained one, so a malformed response would report a checkpoint that never happened.
*/
java/lancedb-core/src/main/java/com/lancedb/LanceDbNamespaceClientBuilder.java
+2 -2
View File
@@ -151,8 +151,8 @@ public class LanceDbNamespaceClientBuilder {
* Build a {@link LanceDbRestClient} for the same endpoint.
*
* <p>Needed only for LanceDB routes that the Lance Namespace specification does not cover — the
* MemWAL write path, reached through {@link LanceDbTableMemWal}. Every other table
* operation belongs on the {@link LanceNamespace} from {@link #build()}.
* RTE write path, reached through {@link LanceDbTableRTE}. Every other table operation belongs on
* the {@link LanceNamespace} from {@link #build()}.
*
* <p>The returned client owns an HTTP connection pool; close it when you are done with it.
*
java/lancedb-core/src/main/java/com/lancedb/LanceDbRestClient.java
+7 -8
View File
@@ -31,9 +31,9 @@ import java.io.UncheckedIOException;
* specification does not cover.
*
* <p>Most table operations reach LanceDB through {@link org.lance.namespace.LanceNamespace}, which
* is generated from the namespace spec. A handful of routes — the MemWAL write path in
* particular — are served by the same endpoint but are not part of that spec, so they are issued
* directly here. See {@link LanceDbTableMemWal}.
* is generated from the namespace spec. A handful of routes — the RTE write path in particular —
* are served by the same endpoint but are not part of that spec, so they are issued directly here.
* See {@link LanceDbTableRTE}.
*
* <p>Obtain one from {@link LanceDbNamespaceClientBuilder#buildRestClient()}.
*/
@@ -50,8 +50,8 @@ public class LanceDbRestClient implements Closeable {
this.apiKey = apiKey;
this.database = database;
// Automatic retries off, deliberately. The default strategy retries 429 and 503 —
// exactly the two statuses LanceDbTableMemWal.checkpointMemWal() acts on — which would
// silently double its explicit retry budget and would also retry compact_memwal in
// exactly the two statuses LanceDbTableRTE.checkpointRTE() acts on — which would
// silently double its explicit retry budget and would also retry compact_rte in
// place, where the loop is designed to fall through to a fresh stats poll instead.
// The checkpoint loop owns the 421/429/503 transitions; the transport must not.
this.http = HttpClients.custom().disableAutomaticRetries().build();
@@ -98,9 +98,8 @@ public class LanceDbRestClient implements Closeable {
/**
* A non-2xx response.
*
* <p>The status is exposed because callers act on it: {@link
* LanceDbTableMemWal#checkpointMemWal()} treats 429 and 503 as retryable and 421 as a lost node
* claim.
* <p>The status is exposed because callers act on it: {@link LanceDbTableRTE#checkpointRTE()}
* treats 429 and 503 as retryable and 421 as a lost node claim.
*/
public static class HttpException extends RuntimeException {
private static final long serialVersionUID = 1L;
java/lancedb-core/src/main/java/com/lancedb/LanceDbTableMemWal.java → java/lancedb-core/src/main/java/com/lancedb/LanceDbTableRTE.java
+58 -58
View File
@@ -22,11 +22,11 @@ import java.util.Optional;
import java.util.OptionalLong;
/**
* The MemWAL write path for one LanceDB Cloud or Enterprise table.
* The RTE write path for one LanceDB Cloud or Enterprise table.
*
* <p>Installing an {@link MemWalWriteSpec} routes {@code mergeInsert} upserts through Lance's
* MemWAL — an LSM-style append — instead of the standard merge path. Rows land in an in-memory
* memtable, freeze into SSTables, and are merged into the base table by compaction.
* <p>Installing an {@link RTEWriteSpec} routes {@code mergeInsert} upserts through Lance's RTE — an
* LSM-style append — instead of the standard merge path. Rows land in an in-memory memtable, freeze
* into SSTables, and are merged into the base table by compaction.
*
* <p>These routes are not part of the Lance Namespace specification, so they are issued directly
* rather than through {@link org.lance.namespace.LanceNamespace}.
@@ -37,23 +37,23 @@ import java.util.OptionalLong;
* .database("your_database_name")
* .buildRestClient();
*
* LanceDbTableMemWal memWal = new LanceDbTableMemWal(client, "my_table");
* memWal.enableMemWal(MemWalWriteSpec.bucket("id", 16));
* LanceDbTableRTE rte = new LanceDbTableRTE(client, "my_table");
* rte.enableRTE(RTEWriteSpec.bucket("id", 16));
* // ... merge_insert traffic ...
* memWal.checkpointMemWal();
* rte.checkpointRTE();
* }</pre>
*/
public class LanceDbTableMemWal {
public class LanceDbTableRTE {
/**
* Interval between {@code get_memwal_stats} polls during a checkpoint. One interval is roughly
* one compaction pass, the granularity at which the answer can change.
* Interval between {@code get_rte_stats} polls during a checkpoint. One interval is roughly one
* compaction pass, the granularity at which the answer can change.
*/
private static final long POLL_INTERVAL_MS = 5_000L;
/**
* Cap on re-issues from {@code flushMemWal} after a 421, so a crash-looping node cannot turn
* flush → compact → 421 → flush into a spin.
* Cap on re-issues from {@code flushRTE} after a 421, so a crash-looping node cannot turn flush →
* compact → 421 → flush into a spin.
*
* <p>Deliberately not shared with {@link #MAX_RETRIES}: a claim that keeps evaporating is a
* broken node, while contention is routine and wants a real budget.
@@ -79,7 +79,7 @@ public class LanceDbTableMemWal {
* @param tableIdentifier The table's full identifier, {@code $}-delimited when it sits inside a
* namespace, such as {@code analytics$events}.
*/
public LanceDbTableMemWal(LanceDbRestClient client, String tableIdentifier) {
public LanceDbTableRTE(LanceDbRestClient client, String tableIdentifier) {
if (client == null) {
throw new IllegalArgumentException("Client cannot be null");
}
@@ -91,46 +91,46 @@ public class LanceDbTableMemWal {
}
/**
* Install an {@link MemWalWriteSpec} on this table, selecting the MemWAL write path for
* future {@code mergeInsert} calls.
* Install an {@link RTEWriteSpec} on this table, selecting the RTE write path for future {@code
* mergeInsert} calls.
*
* <p>All variants require the table to have an unenforced primary key; bucket sharding
* additionally requires it to be the single column being bucketed.
*/
public void enableMemWal(MemWalWriteSpec spec) {
public void enableRTE(RTEWriteSpec spec) {
if (spec == null) {
throw new IllegalArgumentException("Spec cannot be null");
}
client.post(route("enable_memwal"), spec.toRequestBody());
client.post(route("enable_rte"), spec.toRequestBody());
}
/**
* Remove the {@link MemWalWriteSpec} from this table, reverting to the standard {@code
* mergeInsert} write path.
* Remove the {@link RTEWriteSpec} from this table, reverting to the standard {@code mergeInsert}
* write path.
*
* <p>Errors if no spec is currently set.
*/
public void disableMemWal() {
client.post(route("disable_memwal"), null);
public void disableRTE() {
client.post(route("disable_rte"), null);
}
/**
* Read the {@link MemWalWriteSpec} currently installed on this table.
* Read the {@link RTEWriteSpec} currently installed on this table.
*
* <p>Empty when the MemWAL write path is not enabled at all — no spec has been set, or one was
* removed with {@link #disableMemWal()}. That is a different answer from a present spec whose
* {@link MemWalWriteSpec#maintainedIndexes()} is null, which selects indexes automatically.
* <p>Empty when the RTE write path is not enabled at all — no spec has been set, or one was
* removed with {@link #disableRTE()}. That is a different answer from a present spec whose {@link
* RTEWriteSpec#maintainedIndexes()} is null, which selects indexes automatically.
*
* <p>The spec read back is the one that was installed, selection included: null maintains every
* supported index the table has now or gains later, an empty list maintains none, and a non-empty
* list maintains exactly those. All three round-trip.
*/
public Optional<MemWalWriteSpec> getMemWalSpec() {
JsonNode response = client.post(route("get_memwal_spec"), null);
if (response == null || !response.hasNonNull("memwal_spec")) {
public Optional<RTEWriteSpec> getRTESpec() {
JsonNode response = client.post(route("get_rte_spec"), null);
if (response == null || !response.hasNonNull("rte_spec")) {
return Optional.empty();
}
return Optional.of(MemWalWriteSpec.fromJson(response.get("memwal_spec")));
return Optional.of(RTEWriteSpec.fromJson(response.get("rte_spec")));
}
/**
@@ -139,56 +139,56 @@ public class LanceDbTableMemWal {
* <p>Returns once the freeze is committed. Freezing an empty memtable is a no-op, so this is safe
* to call repeatedly.
*/
public void flushMemWal() {
client.post(route("flush_memwal"), null);
public void flushRTE() {
client.post(route("flush_rte"), null);
}
/**
* Trigger a background SSTable compaction pass per table shard.
*
* <p>Returns once the passes are <em>dispatched</em>, not once they finish — watch {@link
* #getMemWalStats}, or use {@link #checkpointMemWal} to wait for convergence.
* #getRTEStats}, or use {@link #checkpointRTE} to wait for convergence.
*/
public void compactMemWal() {
client.post(route("compact_memwal"), null);
public void compactRTE() {
client.post(route("compact_rte"), null);
}
/**
* Read live per-table-shard MemWAL state.
* Read live per-table-shard RTE state.
*
* <p>Answers "how far behind is my fresh tier", "which table shard is hot", and "why is my
* fresh-tier vector search brute-force". Mutates no table state.
*
* <p>Empty only when the MemWAL write path is not enabled — that is, when the server sends an
* absent or null {@code memwal_stats}. A stats object that is present is decoded strictly, and a
* malformed one throws rather than decoding to something empty, because {@link #checkpointMemWal}
* reads convergence out of these numbers and cannot tell a defaulted array from a drained one.
* <p>Empty only when the RTE write path is not enabled — that is, when the server sends an absent
* or null {@code rte_stats}. A stats object that is present is decoded strictly, and a malformed
* one throws rather than decoding to something empty, because {@link #checkpointRTE} reads
* convergence out of these numbers and cannot tell a defaulted array from a drained one.
*
* @param includeSstableRows Also count rows per SSTable. Off by default because each count opens
* an uncached Lance dataset.
* @throws IllegalStateException if the response is absent or does not decode.
*/
public Optional<MemWalStats> getMemWalStats(boolean includeSstableRows) {
public Optional<RTEStats> getRTEStats(boolean includeSstableRows) {
Map<String, Object> body = new LinkedHashMap<String, Object>();
body.put("include_sstable_rows", includeSstableRows);
JsonNode response = client.post(route("get_memwal_stats"), body);
JsonNode response = client.post(route("get_rte_stats"), body);
if (response == null) {
throw new IllegalStateException("get_memwal_stats returned an empty response body");
throw new IllegalStateException("get_rte_stats returned an empty response body");
}
JsonNode stats = response.get("memwal_stats");
JsonNode stats = response.get("rte_stats");
if (stats == null || stats.isNull()) {
return Optional.empty();
}
return Optional.of(MemWalStats.fromJson(stats));
return Optional.of(RTEStats.fromJson(stats));
}
/** Equivalent to {@code getMemWalStats(false)}. */
public Optional<MemWalStats> getMemWalStats() {
return getMemWalStats(false);
/** Equivalent to {@code getRTEStats(false)}. */
public Optional<RTEStats> getRTEStats() {
return getRTEStats(false);
}
/**
* Converge this table's MemWAL write path into its base table.
* Converge this table's RTE write path into its base table.
*
* <p>Freezes once, fixes a target watermark from the resulting SSTables, then triggers compaction
* and polls until those SSTables are gone. The target set is fixed at the start, so sstables
@@ -196,31 +196,31 @@ public class LanceDbTableMemWal {
* load, and what makes it best-effort: it converges the fresh tier as of some instant.
* Idempotent, abandonable at any point, safe on a cadence.
*
* <p>The loop runs here, not on the server: {@link #compactMemWal} dispatches a pass and returns,
* so nothing holds a socket and a client can vanish mid-operation with nothing to reconcile.
* <p>The loop runs here, not on the server: {@link #compactRTE} dispatches a pass and returns, so
* nothing holds a socket and a client can vanish mid-operation with nothing to reconcile.
* Completion is read from generation numbers in the shard manifest — durable state, unlike a
* count in a compact response, which a concurrent write invalidates.
*
* <p>No liveness bound — the caller owns the deadline. The compactor pool is shared across
* tables, so a checkpoint queued behind unrelated work looks exactly like one that is merging.
*/
public void checkpointMemWal() {
public void checkpointRTE() {
for (int reissue = 0; reissue <= MAX_REISSUES; reissue++) {
// The freeze turns everything written before this call into an SSTable, so the
// watermark has to be read after it. Idempotent: freezing an empty memtable is a
// no-op, so a re-issue does not churn empty sstables.
if (issueVoid(this::flushMemWal)) {
if (issueVoid(this::flushRTE)) {
backoff(reissue);
continue;
}
Attempt<Optional<MemWalStats>> stats = issue(() -> getMemWalStats(false));
Attempt<Optional<RTEStats>> stats = issue(() -> getRTEStats(false));
if (stats.lostClaim) {
backoff(reissue);
continue;
}
if (!stats.value.isPresent()) {
// Not WAL-backed; flushMemWal would have errored first but for a race.
// Not WAL-backed; flushRTE would have errored first but for a race.
return;
}
@@ -235,7 +235,7 @@ public class LanceDbTableMemWal {
backoff(reissue);
}
throw new IllegalStateException(
"checkpointMemWal: the owning node kept losing its claim; re-issued from flush the maximum "
"checkpointRTE: the owning node kept losing its claim; re-issued from flush the maximum "
+ "number of times");
}
@@ -246,7 +246,7 @@ public class LanceDbTableMemWal {
*/
private boolean drainToTargets(Map<String, Long> targets) {
while (true) {
Attempt<Optional<MemWalStats>> stats = issue(() -> getMemWalStats(false));
Attempt<Optional<RTEStats>> stats = issue(() -> getRTEStats(false));
if (stats.lostClaim) {
return false;
}
@@ -277,7 +277,7 @@ public class LanceDbTableMemWal {
if (!allCompacting) {
try {
compactMemWal();
compactRTE();
} catch (LanceDbRestClient.HttpException e) {
if (isLostClaim(e)) {
return false;
@@ -296,7 +296,7 @@ public class LanceDbTableMemWal {
}
/** The newest SSTable generation held by each table shard, skipping table shards holding none. */
private static Map<String, Long> newestSstableGenerations(MemWalStats stats) {
private static Map<String, Long> newestSstableGenerations(RTEStats stats) {
Map<String, Long> targets = new HashMap<String, Long>();
for (TableShardStats tableShard : stats.tableShards()) {
OptionalLong newest = tableShard.newestSstableGeneration();
java/lancedb-core/src/main/java/com/lancedb/MemWalStats.java → java/lancedb-core/src/main/java/com/lancedb/RTEStats.java
+7 -7
View File
@@ -20,18 +20,18 @@ import java.util.Collections;
import java.util.List;
/**
* Live per-table-shard MemWAL state, as returned by {@link LanceDbTableMemWal#getMemWalStats()}.
* Live per-table-shard RTE state, as returned by {@link LanceDbTableRTE#getRTEStats()}.
*
* <p>Nothing here is derived: sums and differences (total SSTable bytes, WAL lag) are the caller's
* to compute. There is no "LSM is off" shape — that case is an empty {@link java.util.Optional},
* because a stats object of zeros would read as measurements.
*/
public class MemWalStats {
private static final String CONTEXT = "memwal stats";
public class RTEStats {
private static final String CONTEXT = "rte stats";
private final List<TableShardStats> tableShards;
MemWalStats(List<TableShardStats> tableShards) {
RTEStats(List<TableShardStats> tableShards) {
this.tableShards = Collections.unmodifiableList(tableShards);
}
@@ -40,17 +40,17 @@ public class MemWalStats {
return tableShards;
}
static MemWalStats fromJson(JsonNode node) {
static RTEStats fromJson(JsonNode node) {
JsonFields.requiredObject(node, CONTEXT);
List<TableShardStats> tableShards = new ArrayList<TableShardStats>();
for (JsonNode tableShard : JsonFields.requiredArray(node, "table_shards", CONTEXT)) {
tableShards.add(TableShardStats.fromJson(tableShard));
}
return new MemWalStats(tableShards);
return new RTEStats(tableShards);
}
@Override
public String toString() {
return "MemWalStats{tableShards=" + tableShards + "}";
return "RTEStats{tableShards=" + tableShards + "}";
}
}
java/lancedb-core/src/main/java/com/lancedb/MemWalWriteSpec.java → java/lancedb-core/src/main/java/com/lancedb/RTEWriteSpec.java
+26 -26
View File
@@ -23,22 +23,23 @@ import java.util.List;
import java.util.Map;
/**
* Specification selecting Lance's MemWAL LSM-style write path for {@code mergeInsert}.
* Specification enabling the realtime engine (RTE), which routes {@code mergeInsert} through
* Lance's MemWAL LSM-style write path.
*
* <p>Construct via {@link #bucket}, {@link #identity}, or {@link #unsharded}, then optionally chain
* {@link #withMaintainedIndexes} and {@link #withWriterConfigDefaults}. Install it with {@link
* LanceDbTableMemWal#enableMemWal} and remove it with {@link LanceDbTableMemWal#disableMemWal}.
* LanceDbTableRTE#enableRTE} and remove it with {@link LanceDbTableRTE#disableRTE}.
*
* <p>This is deliberately not {@code org.lance.memwal.InitializeMemWalParams}. That type is Lance's
* own, and its maintained-index default is the opposite of this one: it defaults to maintaining
* <em>nothing</em>, while a fresh spec here maintains <em>every</em> index. It also cannot express
* the null that asks the server to resolve the set.
*/
public class MemWalWriteSpec {
public class RTEWriteSpec {
private static final String CONTEXT = "get_memwal_spec response";
private static final String CONTEXT = "get_rte_spec response";
/** How writes are routed to MemWAL shards. */
/** How writes are routed to RTE shards. */
public enum Sharding {
/** Hash-bucket writes by a scalar column. */
BUCKET("bucket"),
@@ -73,7 +74,7 @@ public class MemWalWriteSpec {
private final List<String> maintainedIndexes;
private final Map<String, String> writerConfigDefaults;
private MemWalWriteSpec(
private RTEWriteSpec(
Sharding sharding,
String column,
Integer numBuckets,
@@ -95,11 +96,11 @@ public class MemWalWriteSpec {
* @param column A non-nested column with a supported scalar type.
* @param numBuckets The number of buckets, in {@code [1, 1024]}.
*/
public static MemWalWriteSpec bucket(String column, int numBuckets) {
public static RTEWriteSpec bucket(String column, int numBuckets) {
if (column == null || column.trim().isEmpty()) {
throw new IllegalArgumentException("Column cannot be null or empty");
}
return new MemWalWriteSpec(
return new RTEWriteSpec(
Sharding.BUCKET, column, numBuckets, null, new HashMap<String, String>());
}
@@ -111,29 +112,28 @@ public class MemWalWriteSpec {
* with a given primary key must always produce the same {@code column} value, or upserts of that
* key can land in different shards and a stale version can win.
*/
public static MemWalWriteSpec identity(String column) {
public static RTEWriteSpec identity(String column) {
if (column == null || column.trim().isEmpty()) {
throw new IllegalArgumentException("Column cannot be null or empty");
}
return new MemWalWriteSpec(
Sharding.IDENTITY, column, null, null, new HashMap<String, String>());
return new RTEWriteSpec(Sharding.IDENTITY, column, null, null, new HashMap<String, String>());
}
/** No sharding — every write goes to a single MemWAL shard — maintaining every index. */
public static MemWalWriteSpec unsharded() {
return new MemWalWriteSpec(Sharding.UNSHARDED, null, null, null, new HashMap<String, String>());
/** No sharding — every write goes to a single RTE shard — maintaining every index. */
public static RTEWriteSpec unsharded() {
return new RTEWriteSpec(Sharding.UNSHARDED, null, null, null, new HashMap<String, String>());
}
/**
* Set the indexes the MemWAL keeps up to date as rows are appended.
* Set the indexes the RTE keeps up to date as rows are appended.
*
* <p>Pass {@code null} — the default for a fresh spec — to maintain every index the table has,
* including ones created later. Pass an empty list to maintain none.
*
* <p>Note that {@code null} and the empty list mean opposite things here.
*/
public MemWalWriteSpec withMaintainedIndexes(List<String> maintainedIndexes) {
return new MemWalWriteSpec(
public RTEWriteSpec withMaintainedIndexes(List<String> maintainedIndexes) {
return new RTEWriteSpec(
sharding,
column,
numBuckets,
@@ -150,11 +150,11 @@ public class MemWalWriteSpec {
* max_unflushed_memtable_bytes}, and {@code enable_memtable}. Duration knobs carry an {@code _ms}
* suffix, such as {@code max_wal_flush_interval_ms}.
*/
public MemWalWriteSpec withWriterConfigDefaults(Map<String, String> writerConfigDefaults) {
public RTEWriteSpec withWriterConfigDefaults(Map<String, String> writerConfigDefaults) {
if (writerConfigDefaults == null) {
throw new IllegalArgumentException("writerConfigDefaults cannot be null");
}
return new MemWalWriteSpec(
return new RTEWriteSpec(
sharding,
column,
numBuckets,
@@ -178,7 +178,7 @@ public class MemWalWriteSpec {
}
/**
* Which indexes the MemWAL maintains.
* Which indexes the RTE maintains.
*
* <p>Null selects automatically: every supported index the table has now or gains later. An empty
* list maintains none. A non-empty list maintains exactly those.
@@ -192,7 +192,7 @@ public class MemWalWriteSpec {
return Collections.unmodifiableMap(writerConfigDefaults);
}
/** Render this spec as the {@code enable_memwal} request body. */
/** Render this spec as the {@code enable_rte} request body. */
Map<String, Object> toRequestBody() {
Map<String, Object> shardingBody = new LinkedHashMap<String, Object>();
shardingBody.put("mode", sharding.wireName());
@@ -212,15 +212,15 @@ public class MemWalWriteSpec {
}
/**
* Rebuild a spec from a {@code get_memwal_spec} response body.
* Rebuild a spec from a {@code get_rte_spec} response body.
*
* <p>A null {@code maintained_indexes} selects every index the table has; an empty list selects
* none.
*/
static MemWalWriteSpec fromJson(JsonNode node) {
static RTEWriteSpec fromJson(JsonNode node) {
JsonNode shardingNode = node.get("sharding");
if (shardingNode == null || shardingNode.get("mode") == null) {
throw new IllegalStateException("get_memwal_spec response has no sharding mode");
throw new IllegalStateException("get_rte_spec response has no sharding mode");
}
Sharding sharding = Sharding.fromWireName(shardingNode.get("mode").asText());
@@ -253,12 +253,12 @@ public class MemWalWriteSpec {
.forEachRemaining(name -> defaults.put(name, defaultsNode.get(name).asText()));
}
return new MemWalWriteSpec(sharding, column, numBuckets, maintainedIndexes, defaults);
return new RTEWriteSpec(sharding, column, numBuckets, maintainedIndexes, defaults);
}
@Override
public String toString() {
return "MemWalWriteSpec{sharding="
return "RTEWriteSpec{sharding="
+ sharding
+ ", column="
+ column
java/lancedb-core/src/test/java/com/lancedb/LanceDbTableMemWalTest.java → java/lancedb-core/src/test/java/com/lancedb/LanceDbTableRTETest.java
+129 -133
View File
@@ -41,17 +41,17 @@ import java.util.concurrent.ConcurrentHashMap;
import static org.junit.jupiter.api.Assertions.*;
/**
* Unit tests for the MemWAL LSM routes, run against a scripted local HTTP server.
* Unit tests for the RTE LSM routes, run against a scripted local HTTP server.
*
* <p>The wire assertions mirror the Rust mocked-endpoint tests in {@code
* rust/lancedb/src/remote/table.rs}, which are the contract these routes have to match.
*/
public class LanceDbTableMemWalTest {
public class LanceDbTableRTETest {
private static final ObjectMapper MAPPER = new ObjectMapper();
private HttpServer server;
private LanceDbRestClient client;
private LanceDbTableMemWal memWal;
private LanceDbTableRTE rte;
private final List<String> requestPaths = Collections.synchronizedList(new ArrayList<String>());
private final List<String> requestBodies = Collections.synchronizedList(new ArrayList<String>());
@@ -101,7 +101,7 @@ public class LanceDbTableMemWalTest {
.database("test-db")
.endpoint("http://127.0.0.1:" + server.getAddress().getPort())
.buildRestClient();
memWal = new LanceDbTableMemWal(client, "my_table");
rte = new LanceDbTableRTE(client, "my_table");
}
@AfterEach
@@ -115,12 +115,12 @@ public class LanceDbTableMemWalTest {
// ===========================================================================
@Test
public void testEnableMemWalUnsharded() throws Exception {
enqueue("enable_memwal", 200, "");
public void testEnableRTEUnsharded() throws Exception {
enqueue("enable_rte", 200, "");
memWal.enableMemWal(MemWalWriteSpec.unsharded());
rte.enableRTE(RTEWriteSpec.unsharded());
assertEquals("/v1/table/my_table/enable_memwal/", requestPaths.get(0));
assertEquals("/v1/table/my_table/enable_rte/", requestPaths.get(0));
JsonNode body = MAPPER.readTree(requestBodies.get(0));
assertEquals("unsharded", body.get("sharding").get("mode").asText());
assertFalse(body.get("sharding").has("column"));
@@ -128,11 +128,10 @@ public class LanceDbTableMemWalTest {
}
@Test
public void testEnableMemWalBucket() throws Exception {
enqueue("enable_memwal", 200, "");
public void testEnableRTEBucket() throws Exception {
enqueue("enable_rte", 200, "");
memWal.enableMemWal(
MemWalWriteSpec.bucket("id", 16).withMaintainedIndexes(Arrays.asList("id_idx")));
rte.enableRTE(RTEWriteSpec.bucket("id", 16).withMaintainedIndexes(Arrays.asList("id_idx")));
JsonNode body = MAPPER.readTree(requestBodies.get(0));
assertEquals("bucket", body.get("sharding").get("mode").asText());
@@ -143,10 +142,10 @@ public class LanceDbTableMemWalTest {
}
@Test
public void testEnableMemWalIdentity() throws Exception {
enqueue("enable_memwal", 200, "");
public void testEnableRTEIdentity() throws Exception {
enqueue("enable_rte", 200, "");
memWal.enableMemWal(MemWalWriteSpec.identity("tenant"));
rte.enableRTE(RTEWriteSpec.identity("tenant"));
JsonNode body = MAPPER.readTree(requestBodies.get(0));
assertEquals("identity", body.get("sharding").get("mode").asText());
@@ -160,55 +159,54 @@ public class LanceDbTableMemWalTest {
*/
@Test
public void testMaintainedIndexesNullAndEmptyAreDistinctOnTheWire() throws Exception {
enqueue("enable_memwal", 200, "");
enqueue("enable_rte", 200, "");
memWal.enableMemWal(MemWalWriteSpec.unsharded());
rte.enableRTE(RTEWriteSpec.unsharded());
JsonNode fresh = MAPPER.readTree(requestBodies.get(0));
assertTrue(fresh.has("maintained_indexes"), "the key must be present");
assertTrue(fresh.get("maintained_indexes").isNull(), "a fresh spec sends null, not []");
memWal.enableMemWal(
MemWalWriteSpec.unsharded().withMaintainedIndexes(Collections.<String>emptyList()));
rte.enableRTE(RTEWriteSpec.unsharded().withMaintainedIndexes(Collections.<String>emptyList()));
JsonNode none = MAPPER.readTree(requestBodies.get(1));
assertTrue(none.get("maintained_indexes").isArray());
assertEquals(0, none.get("maintained_indexes").size());
}
@Test
public void testEnableMemWalWriterConfigDefaults() throws Exception {
enqueue("enable_memwal", 200, "");
public void testEnableRTEWriterConfigDefaults() throws Exception {
enqueue("enable_rte", 200, "");
Map<String, String> defaults = new HashMap<String, String>();
defaults.put("max_memtable_rows", "50000");
memWal.enableMemWal(MemWalWriteSpec.unsharded().withWriterConfigDefaults(defaults));
rte.enableRTE(RTEWriteSpec.unsharded().withWriterConfigDefaults(defaults));
JsonNode body = MAPPER.readTree(requestBodies.get(0));
assertEquals("50000", body.get("writer_config_defaults").get("max_memtable_rows").asText());
}
@Test
public void testDisableMemWal() {
enqueue("disable_memwal", 200, "");
public void testDisableRTE() {
enqueue("disable_rte", 200, "");
memWal.disableMemWal();
rte.disableRTE();
assertEquals("/v1/table/my_table/disable_memwal/", requestPaths.get(0));
assertEquals("/v1/table/my_table/disable_rte/", requestPaths.get(0));
assertEquals("", requestBodies.get(0));
}
@Test
public void testGetMemWalSpec() {
public void testGetRTESpec() {
enqueue(
"get_memwal_spec",
"get_rte_spec",
200,
"{\"memwal_spec\":{\"sharding\":{\"mode\":\"bucket\",\"column\":\"id\","
"{\"rte_spec\":{\"sharding\":{\"mode\":\"bucket\",\"column\":\"id\","
+ "\"num_buckets\":16},\"maintained_indexes\":[\"id_idx\"],"
+ "\"writer_config_defaults\":{\"durable_write\":\"true\"}}}");
Optional<MemWalWriteSpec> spec = memWal.getMemWalSpec();
Optional<RTEWriteSpec> spec = rte.getRTESpec();
assertTrue(spec.isPresent());
assertEquals(MemWalWriteSpec.Sharding.BUCKET, spec.get().sharding());
assertEquals(RTEWriteSpec.Sharding.BUCKET, spec.get().sharding());
assertEquals("id", spec.get().column());
assertEquals(Integer.valueOf(16), spec.get().numBuckets());
assertEquals(Arrays.asList("id_idx"), spec.get().maintainedIndexes());
@@ -217,28 +215,28 @@ public class LanceDbTableMemWalTest {
/** A null selection, every index the table has, reads back as null rather than an empty list. */
@Test
public void testGetMemWalSpecRoundTripsEveryIndexSelection() {
public void testGetRTESpecRoundTripsEveryIndexSelection() {
enqueue(
"get_memwal_spec",
"get_rte_spec",
200,
"{\"memwal_spec\":{\"sharding\":{\"mode\":\"unsharded\"},"
"{\"rte_spec\":{\"sharding\":{\"mode\":\"unsharded\"},"
+ "\"maintained_indexes\":null,\"writer_config_defaults\":{}}}");
Optional<MemWalWriteSpec> spec = memWal.getMemWalSpec();
Optional<RTEWriteSpec> spec = rte.getRTESpec();
assertTrue(spec.isPresent());
assertNull(spec.get().maintainedIndexes());
}
@Test
public void testGetMemWalSpecRoundTripsNoIndexSelection() {
public void testGetRTESpecRoundTripsNoIndexSelection() {
enqueue(
"get_memwal_spec",
"get_rte_spec",
200,
"{\"memwal_spec\":{\"sharding\":{\"mode\":\"unsharded\"},"
"{\"rte_spec\":{\"sharding\":{\"mode\":\"unsharded\"},"
+ "\"maintained_indexes\":[],\"writer_config_defaults\":{}}}");
Optional<MemWalWriteSpec> spec = memWal.getMemWalSpec();
Optional<RTEWriteSpec> spec = rte.getRTESpec();
assertTrue(spec.isPresent());
assertEquals(Collections.emptyList(), spec.get().maintainedIndexes());
@@ -249,32 +247,32 @@ public class LanceDbTableMemWalTest {
* rejected: reading a string or an object as null would turn a malformed response into a policy.
*/
@Test
public void testGetMemWalSpecRejectsANonArrayMaintainedIndexes() {
public void testGetRTESpecRejectsANonArrayMaintainedIndexes() {
enqueue(
"get_memwal_spec",
"get_rte_spec",
200,
"{\"memwal_spec\":{\"sharding\":{\"mode\":\"unsharded\"},"
"{\"rte_spec\":{\"sharding\":{\"mode\":\"unsharded\"},"
+ "\"maintained_indexes\":\"id_idx\",\"writer_config_defaults\":{}}}");
assertThrows(IllegalStateException.class, () -> memWal.getMemWalSpec());
assertThrows(IllegalStateException.class, () -> rte.getRTESpec());
}
@Test
public void testGetMemWalSpecRejectsNonStringIndexNames() {
public void testGetRTESpecRejectsNonStringIndexNames() {
enqueue(
"get_memwal_spec",
"get_rte_spec",
200,
"{\"memwal_spec\":{\"sharding\":{\"mode\":\"unsharded\"},"
"{\"rte_spec\":{\"sharding\":{\"mode\":\"unsharded\"},"
+ "\"maintained_indexes\":[7],\"writer_config_defaults\":{}}}");
assertThrows(IllegalStateException.class, () -> memWal.getMemWalSpec());
assertThrows(IllegalStateException.class, () -> rte.getRTESpec());
}
@Test
public void testGetMemWalSpecAbsent() {
enqueue("get_memwal_spec", 200, "{\"memwal_spec\":null}");
public void testGetRTESpecAbsent() {
enqueue("get_rte_spec", 200, "{\"rte_spec\":null}");
assertFalse(memWal.getMemWalSpec().isPresent());
assertFalse(rte.getRTESpec().isPresent());
}
// ===========================================================================
@@ -282,12 +280,12 @@ public class LanceDbTableMemWalTest {
// ===========================================================================
@Test
public void testGetMemWalStats() throws Exception {
enqueue("get_memwal_stats", 200, stats(tableShard("shard-0", false, 7L, 8L)));
public void testGetRTEStats() throws Exception {
enqueue("get_rte_stats", 200, stats(tableShard("shard-0", false, 7L, 8L)));
Optional<MemWalStats> got = memWal.getMemWalStats(true);
Optional<RTEStats> got = rte.getRTEStats(true);
assertEquals("/v1/table/my_table/get_memwal_stats/", requestPaths.get(0));
assertEquals("/v1/table/my_table/get_rte_stats/", requestPaths.get(0));
assertTrue(MAPPER.readTree(requestBodies.get(0)).get("include_sstable_rows").asBoolean());
assertTrue(got.isPresent());
TableShardStats decoded = got.get().tableShards().get(0);
@@ -305,18 +303,18 @@ public class LanceDbTableMemWalTest {
/** The optional fields decode when the server does send them. */
@Test
public void testGetMemWalStatsDecodesOptionalFields() {
public void testGetRTEStatsDecodesOptionalFields() {
enqueue(
"get_memwal_stats",
"get_rte_stats",
200,
"{\"memwal_stats\":{\"table_shards\":[{\"shard_id\":\"shard-0\",\"status\":\"Active\","
"{\"rte_stats\":{\"table_shards\":[{\"shard_id\":\"shard-0\",\"status\":\"Active\","
+ "\"writer_epoch\":1,\"manifest_version\":2,\"current_generation\":9,"
+ "\"replay_after_wal_entry_position\":3,\"wal_entry_position_last_seen\":11,"
+ "\"sstables\":[{\"generation\":7,\"bytes\":1024,\"rows\":42}],"
+ "\"compacting\":true,\"memtables\":[{\"generation\":8,\"rows\":5,"
+ "\"bytes\":64,\"batches\":2,\"indexes\":[\"id_idx\"]}]}]}}");
TableShardStats decoded = memWal.getMemWalStats(true).get().tableShards().get(0);
TableShardStats decoded = rte.getRTEStats(true).get().tableShards().get(0);
assertEquals(3, decoded.replayAfterWalEntryPosition());
assertEquals(11, decoded.walEntryPositionLastSeen());
@@ -332,17 +330,17 @@ public class LanceDbTableMemWalTest {
}
@Test
public void testGetMemWalStatsAbsentWhenMemWalDisabled() {
enqueue("get_memwal_stats", 200, "{\"memwal_stats\":null}");
public void testGetRTEStatsAbsentWhenRTEDisabled() {
enqueue("get_rte_stats", 200, "{\"rte_stats\":null}");
assertFalse(memWal.getMemWalStats().isPresent());
assertFalse(rte.getRTEStats().isPresent());
}
@Test
public void testGetMemWalStatsDefaultsToExcludingGenerationRows() throws Exception {
enqueue("get_memwal_stats", 200, stats());
public void testGetRTEStatsDefaultsToExcludingGenerationRows() throws Exception {
enqueue("get_rte_stats", 200, stats());
memWal.getMemWalStats();
rte.getRTEStats();
assertFalse(MAPPER.readTree(requestBodies.get(0)).get("include_sstable_rows").asBoolean());
}
@@ -353,22 +351,22 @@ public class LanceDbTableMemWalTest {
@Test
public void testFlushAndCompactRoutes() {
enqueue("flush_memwal", 200, "");
enqueue("compact_memwal", 200, "");
enqueue("flush_rte", 200, "");
enqueue("compact_rte", 200, "");
memWal.flushMemWal();
memWal.compactMemWal();
rte.flushRTE();
rte.compactRTE();
assertEquals("/v1/table/my_table/flush_memwal/", requestPaths.get(0));
assertEquals("/v1/table/my_table/compact_memwal/", requestPaths.get(1));
assertEquals("/v1/table/my_table/flush_rte/", requestPaths.get(0));
assertEquals("/v1/table/my_table/compact_rte/", requestPaths.get(1));
}
@Test
public void testHttpErrorCarriesStatus() {
enqueue("flush_memwal", 404, "no such table");
enqueue("flush_rte", 404, "no such table");
LanceDbRestClient.HttpException e =
assertThrows(LanceDbRestClient.HttpException.class, () -> memWal.flushMemWal());
assertThrows(LanceDbRestClient.HttpException.class, () -> rte.flushRTE());
assertEquals(404, e.statusCode());
}
@@ -377,98 +375,97 @@ public class LanceDbTableMemWalTest {
// ===========================================================================
@Test
public void testCheckpointReturnsWhenMemWalDisabled() {
enqueue("flush_memwal", 200, "");
enqueue("get_memwal_stats", 200, "{\"memwal_stats\":null}");
public void testCheckpointReturnsWhenRTEDisabled() {
enqueue("flush_rte", 200, "");
enqueue("get_rte_stats", 200, "{\"rte_stats\":null}");
memWal.checkpointMemWal();
rte.checkpointRTE();
assertEquals(0, countCalls("compact_memwal"), "nothing to compact when the LSM path is off");
assertEquals(0, countCalls("compact_rte"), "nothing to compact when the LSM path is off");
}
@Test
public void testCheckpointReturnsWhenNoGenerationsOutstanding() {
enqueue("flush_memwal", 200, "");
enqueue("flush_rte", 200, "");
// A table shard with no SSTables yields no target, so the drain never starts.
enqueue("get_memwal_stats", 200, stats(tableShard("shard-0", false)));
enqueue("get_rte_stats", 200, stats(tableShard("shard-0", false)));
memWal.checkpointMemWal();
rte.checkpointRTE();
assertEquals(0, countCalls("compact_memwal"));
assertEquals(0, countCalls("compact_rte"));
}
@Test
public void testCheckpointConvergesOnceTargetGenerationsAreGone() {
enqueue("flush_memwal", 200, "");
enqueue("flush_rte", 200, "");
// Watermark read: shard-0 holds sstables 7 and 8, so target = 8.
enqueue("get_memwal_stats", 200, stats(tableShard("shard-0", false, 7L, 8L)));
enqueue("get_rte_stats", 200, stats(tableShard("shard-0", false, 7L, 8L)));
// First drain poll: both still outstanding, nothing compacting -> dispatch a pass.
enqueue("get_memwal_stats", 200, stats(tableShard("shard-0", false, 7L, 8L)));
enqueue("get_rte_stats", 200, stats(tableShard("shard-0", false, 7L, 8L)));
// Second drain poll: drained past the target -> done.
enqueue("get_memwal_stats", 200, stats(tableShard("shard-0", false, 9L)));
enqueue("compact_memwal", 200, "");
enqueue("get_rte_stats", 200, stats(tableShard("shard-0", false, 9L)));
enqueue("compact_rte", 200, "");
memWal.checkpointMemWal();
rte.checkpointRTE();
assertEquals(1, countCalls("compact_memwal"), "one pass dispatched");
assertEquals(3, countCalls("get_memwal_stats"), "watermark read plus two drain polls");
assertEquals(1, countCalls("compact_rte"), "one pass dispatched");
assertEquals(3, countCalls("get_rte_stats"), "watermark read plus two drain polls");
}
@Test
public void testCheckpointDoesNotPileOnWhileEveryTargetBucketIsCompacting() {
enqueue("flush_memwal", 200, "");
enqueue("get_memwal_stats", 200, stats(tableShard("shard-0", true, 4L)));
enqueue("flush_rte", 200, "");
enqueue("get_rte_stats", 200, stats(tableShard("shard-0", true, 4L)));
// Still compacting on the first poll, so no pass is dispatched; then it drains.
enqueue("get_memwal_stats", 200, stats(tableShard("shard-0", true, 4L)));
enqueue("get_memwal_stats", 200, stats(tableShard("shard-0", false, 5L)));
enqueue("get_rte_stats", 200, stats(tableShard("shard-0", true, 4L)));
enqueue("get_rte_stats", 200, stats(tableShard("shard-0", false, 5L)));
memWal.checkpointMemWal();
rte.checkpointRTE();
assertEquals(0, countCalls("compact_memwal"), "a latched table shard is left alone");
assertEquals(0, countCalls("compact_rte"), "a latched table shard is left alone");
}
@Test
public void testCheckpointRetriesFromFlushAfterLostClaim() {
// 421 on the watermark read: the node lost its claim, so the whole thing restarts
// from flush rather than retrying the read in place.
enqueue("flush_memwal", 200, "");
enqueue("get_memwal_stats", 421, "no claim");
enqueue("get_memwal_stats", 200, stats(tableShard("shard-0", false)));
enqueue("flush_rte", 200, "");
enqueue("get_rte_stats", 421, "no claim");
enqueue("get_rte_stats", 200, stats(tableShard("shard-0", false)));
memWal.checkpointMemWal();
rte.checkpointRTE();
assertEquals(2, countCalls("flush_memwal"), "re-issued from flush");
assertEquals(2, countCalls("flush_rte"), "re-issued from flush");
}
@Test
public void testCheckpointRetriesRetryableStatusInPlace() {
enqueue("flush_memwal", 429, "latch held");
enqueue("flush_memwal", 200, "");
enqueue("get_memwal_stats", 200, stats(tableShard("shard-0", false)));
enqueue("flush_rte", 429, "latch held");
enqueue("flush_rte", 200, "");
enqueue("get_rte_stats", 200, stats(tableShard("shard-0", false)));
memWal.checkpointMemWal();
rte.checkpointRTE();
assertEquals(2, countCalls("flush_memwal"), "429 retried in place, not re-issued");
assertEquals(2, countCalls("flush_rte"), "429 retried in place, not re-issued");
}
@Test
public void testCheckpointPropagatesTerminalStatus() {
enqueue("flush_memwal", 400, "bad request");
enqueue("flush_rte", 400, "bad request");
LanceDbRestClient.HttpException e =
assertThrows(LanceDbRestClient.HttpException.class, () -> memWal.checkpointMemWal());
assertThrows(LanceDbRestClient.HttpException.class, () -> rte.checkpointRTE());
assertEquals(400, e.statusCode());
assertEquals(1, countCalls("flush_memwal"), "a terminal status is not retried");
assertEquals(1, countCalls("flush_rte"), "a terminal status is not retried");
}
@Test
public void testCheckpointGivesUpAfterRepeatedLostClaims() {
enqueue("flush_memwal", 421, "no claim");
enqueue("flush_rte", 421, "no claim");
IllegalStateException e =
assertThrows(IllegalStateException.class, () -> memWal.checkpointMemWal());
IllegalStateException e = assertThrows(IllegalStateException.class, () -> rte.checkpointRTE());
assertTrue(e.getMessage().contains("kept losing its claim"), e.getMessage());
assertEquals(4, countCalls("flush_memwal"), "the initial attempt plus MAX_REISSUES");
assertEquals(4, countCalls("flush_rte"), "the initial attempt plus MAX_REISSUES");
}
// ===========================================================================
@@ -478,24 +475,24 @@ public class LanceDbTableMemWalTest {
/**
* A stats payload that does not decode must fail closed. Every one of these bodies used to be
* read as "no table shards", which is indistinguishable from a drained table, so {@code
* checkpointMemWal} reported convergence for a checkpoint that never ran.
* checkpointRTE} reported convergence for a checkpoint that never ran.
*/
@Test
public void testCheckpointRejectsMalformedStats() {
Map<String, String> malformed = new LinkedHashMap<String, String>();
malformed.put("no response body at all", "");
malformed.put("stats object with no table shards", "{\"memwal_stats\":{}}");
malformed.put("stats object with no table shards", "{\"rte_stats\":{}}");
malformed.put(
"table shard missing its required fields", "{\"memwal_stats\":{\"table_shards\":[{}]}}");
"table shard missing its required fields", "{\"rte_stats\":{\"table_shards\":[{}]}}");
malformed.put(
"table shard missing sstables",
"{\"memwal_stats\":{\"table_shards\":[{\"shard_id\":\"shard-0\",\"status\":\"Active\","
"{\"rte_stats\":{\"table_shards\":[{\"shard_id\":\"shard-0\",\"status\":\"Active\","
+ "\"writer_epoch\":1,\"manifest_version\":2,\"current_generation\":9,"
+ "\"replay_after_wal_entry_position\":0,\"wal_entry_position_last_seen\":0,"
+ "\"compacting\":false}]}}");
malformed.put(
"generation with a non-numeric generation number",
"{\"memwal_stats\":{\"table_shards\":[{\"shard_id\":\"shard-0\",\"status\":\"Active\","
"{\"rte_stats\":{\"table_shards\":[{\"shard_id\":\"shard-0\",\"status\":\"Active\","
+ "\"writer_epoch\":1,\"manifest_version\":2,\"current_generation\":9,"
+ "\"replay_after_wal_entry_position\":0,\"wal_entry_position_last_seen\":0,"
+ "\"sstables\":[{\"generation\":\"7\",\"bytes\":1024}],"
@@ -503,25 +500,25 @@ public class LanceDbTableMemWalTest {
for (Map.Entry<String, String> each : malformed.entrySet()) {
setUpFresh();
enqueue("flush_memwal", 200, "");
enqueue("get_memwal_stats", 200, each.getValue());
enqueue("flush_rte", 200, "");
enqueue("get_rte_stats", 200, each.getValue());
assertThrows(
IllegalStateException.class,
() -> memWal.checkpointMemWal(),
() -> rte.checkpointRTE(),
each.getKey() + " must not report convergence");
}
}
/** The one shape that legitimately means "this table has no MemWAL write path". */
/** The one shape that legitimately means "this table has no RTE write path". */
@Test
public void testCheckpointTreatsNullStatsAsNotWalBacked() {
enqueue("flush_memwal", 200, "");
enqueue("get_memwal_stats", 200, "{\"memwal_stats\":null}");
enqueue("flush_rte", 200, "");
enqueue("get_rte_stats", 200, "{\"rte_stats\":null}");
memWal.checkpointMemWal();
rte.checkpointRTE();
assertEquals(1, countCalls("get_memwal_stats"));
assertEquals(1, countCalls("get_rte_stats"));
}
// ===========================================================================
@@ -531,19 +528,18 @@ public class LanceDbTableMemWalTest {
/**
* The transport must not retry on the checkpoint loop's behalf. Apache HttpClient's default
* strategy retries exactly 429 and 503 — the two statuses {@code isRetryable} owns — which
* doubled every budget here and also retried {@code compact_memwal} in place, where the loop is
* doubled every budget here and also retried {@code compact_rte} in place, where the loop is
* built to fall through to a fresh stats poll instead.
*/
@Test
public void testCheckpointRetryBudgetIsNotDoubledByTheTransport() {
enqueue("flush_memwal", 429, "latch held");
enqueue("flush_rte", 429, "latch held");
LanceDbRestClient.HttpException e =
assertThrows(LanceDbRestClient.HttpException.class, () -> memWal.checkpointMemWal());
assertThrows(LanceDbRestClient.HttpException.class, () -> rte.checkpointRTE());
assertEquals(429, e.statusCode(), "the exhausted budget propagates the last error as itself");
assertEquals(
9, countCalls("flush_memwal"), "the initial request plus MAX_RETRIES, and no more");
assertEquals(9, countCalls("flush_rte"), "the initial request plus MAX_RETRIES, and no more");
}
// ===========================================================================
@@ -558,9 +554,9 @@ public class LanceDbTableMemWalTest {
return numbers;
}
/** Build a {@code memwal_stats} response body from table shard fragments. */
/** Build a {@code rte_stats} response body from table shard fragments. */
private static String stats(String... tableShards) {
return "{\"memwal_stats\":{\"table_shards\":[" + String.join(",", tableShards) + "]}}";
return "{\"rte_stats\":{\"table_shards\":[" + String.join(",", tableShards) + "]}}";
}
private static String tableShard(String shardId, boolean compacting, Long... sstables) {
@@ -600,7 +596,7 @@ public class LanceDbTableMemWalTest {
return requestPaths.stream().filter(path -> operationOf(path).equals(operation)).count();
}
/** {@code /v1/table/my_table/flush_memwal/} -> {@code flush_memwal}. */
/** {@code /v1/table/my_table/flush_rte/} -> {@code flush_rte}. */
private static String operationOf(String path) {
String[] segments = path.split("/");
return segments.length == 0 ? "" : segments[segments.length - 1];
nodejs/__test__/table.test.ts
+51 -53
View File
@@ -642,11 +642,11 @@ describe.each(arrowVersions)(
);
});
it("should expose useMemWal on takeRowIds as the base-only escape hatch", async () => {
it("should expose useRTE on takeRowIds as the base-only escape hatch", async () => {
await table.add([{ id: 1 }, { id: 2 }, { id: 3 }]);
// useMemWal(false) is reachable on TakeQuery (the escape hatch for MemWAL tables,
// useRTE(false) is reachable on TakeQuery (the escape hatch for RTE tables,
// where take-by-row-id auto-routes to the LSM scanner and is rejected).
const res = await table.takeRowIds([0, 2]).useMemWal(false).toArray();
const res = await table.takeRowIds([0, 2]).useRTE(false).toArray();
expect(res.map((r) => r.id)).toEqual([1, 3]);
});
@@ -4998,7 +4998,7 @@ describe("setUnenforcedPrimaryKey", () => {
});
});
describe("enableMemWal / disableMemWal", () => {
describe("enableRTE / disableRTE", () => {
let tmpDir: tmp.DirResult;
beforeEach(() => {
@@ -5018,16 +5018,16 @@ describe("enableMemWal / disableMemWal", () => {
const table = await makeTable(conn);
await table.setUnenforcedPrimaryKey("id");
await table.enableMemWal({
await table.enableRTE({
specType: "bucket",
column: "id",
numBuckets: 4,
});
await table.disableMemWal();
await table.disableRTE();
// A second unset errors — there is no spec left to remove.
await expect(table.disableMemWal()).rejects.toThrow();
await expect(table.disableRTE()).rejects.toThrow();
// A fresh spec can be installed after unset.
await table.enableMemWal({
await table.enableRTE({
specType: "bucket",
column: "id",
numBuckets: 8,
@@ -5039,8 +5039,8 @@ describe("enableMemWal / disableMemWal", () => {
const table = await makeTable(conn);
await table.setUnenforcedPrimaryKey("id");
await table.enableMemWal({ specType: "unsharded" });
await table.disableMemWal();
await table.enableRTE({ specType: "unsharded" });
await table.disableRTE();
});
it("installs an identity spec", async () => {
@@ -5048,8 +5048,8 @@ describe("enableMemWal / disableMemWal", () => {
const table = await makeTable(conn);
await table.setUnenforcedPrimaryKey("id");
await table.enableMemWal({ specType: "identity", column: "id" });
await table.disableMemWal();
await table.enableRTE({ specType: "identity", column: "id" });
await table.disableRTE();
});
it("rejects an invalid spec", async () => {
@@ -5059,7 +5059,7 @@ describe("enableMemWal / disableMemWal", () => {
await table.setUnenforcedPrimaryKey("id");
// num_buckets out of range.
await expect(
table.enableMemWal({
table.enableRTE({
specType: "bucket",
column: "id",
numBuckets: 0,
@@ -5067,7 +5067,7 @@ describe("enableMemWal / disableMemWal", () => {
).rejects.toThrow();
// Column mismatch.
await expect(
table.enableMemWal({
table.enableRTE({
specType: "bucket",
column: "missing",
numBuckets: 4,
@@ -5075,13 +5075,13 @@ describe("enableMemWal / disableMemWal", () => {
).rejects.toThrow();
});
it("reads back the installed spec via getMemWalSpec", async () => {
it("reads back the installed spec via getRTESpec", async () => {
const conn = await connect(tmpDir.name);
const table = await makeTable(conn);
await table.setUnenforcedPrimaryKey("id");
// Nothing installed yet.
expect(await table.getMemWalSpec()).toBeUndefined();
expect(await table.getRTESpec()).toBeUndefined();
// A real scalar index is needed to name it as a maintained index.
await table.add([{ id: 1 }, { id: 2 }, { id: 3 }]);
@@ -5092,14 +5092,14 @@ describe("enableMemWal / disableMemWal", () => {
// defaults. Lance writer-config keys are canonically snake_case.
// biome-ignore lint/style/useNamingConvention: Lance writer-config keys are snake_case
const writerConfigDefaults = { durable_write: "false" };
await table.enableMemWal({
await table.enableRTE({
specType: "bucket",
column: "id",
numBuckets: 4,
maintainedIndexes: [indexName],
writerConfigDefaults,
});
const spec = await table.getMemWalSpec();
const spec = await table.getRTESpec();
expect(spec).toBeDefined();
expect(spec?.specType).toBe("bucket");
expect(spec?.column).toBe("id");
@@ -5108,19 +5108,19 @@ describe("enableMemWal / disableMemWal", () => {
expect(spec?.writerConfigDefaults).toEqual(writerConfigDefaults);
// After unset, undefined again.
await table.disableMemWal();
expect(await table.getMemWalSpec()).toBeUndefined();
await table.disableRTE();
expect(await table.getRTESpec()).toBeUndefined();
// Identity round-trips (column recovered from the schema).
await table.enableMemWal({ specType: "identity", column: "id" });
const identity = await table.getMemWalSpec();
await table.enableRTE({ specType: "identity", column: "id" });
const identity = await table.getRTESpec();
expect(identity?.specType).toBe("identity");
expect(identity?.column).toBe("id");
await table.disableMemWal();
await table.disableRTE();
// Unsharded round-trips (no routing column).
await table.enableMemWal({ specType: "unsharded" });
const unsharded = await table.getMemWalSpec();
await table.enableRTE({ specType: "unsharded" });
const unsharded = await table.getRTESpec();
expect(unsharded?.specType).toBe("unsharded");
expect(unsharded?.column).toBeFalsy();
});
@@ -5149,7 +5149,7 @@ describe("LSM merge insert", () => {
]);
await table.setUnenforcedPrimaryKey("id");
// numBuckets = 1: every row routes to the single bucket.
await table.enableMemWal({
await table.enableRTE({
specType: "bucket",
column: "id",
numBuckets: 1,
@@ -5169,22 +5169,22 @@ describe("LSM merge insert", () => {
{ id: "c", value: 3 },
{ id: "d", value: 4 },
]);
// LSM path: rows go to the MemWAL, so only numRows is populated.
// LSM path: rows go to the RTE, so only numRows is populated.
expect(res.numRows).toBe(2);
expect(res.version).toBe(0);
expect(res.numInsertedRows).toBe(0);
await table.closeMemWalWriters();
await table.closeRTEWriters();
});
it("falls back to the standard path with useMemWal(false)", async () => {
it("falls back to the standard path with useRTE(false)", async () => {
const conn = await connect(tmpDir.name);
const table = await bucketTable(conn);
const res = await table
.mergeInsert("id")
.whenNotMatchedInsertAll()
.useMemWal(false)
.useRTE(false)
.execute([
{ id: "b", value: 9 },
{ id: "e", value: 5 },
@@ -5207,7 +5207,7 @@ describe("LSM merge insert", () => {
expect(res.numRows).toBe(1);
});
it("rejects a non-upsert merge under a MemWAL spec", async () => {
it("rejects a non-upsert merge under an RTE spec", async () => {
const conn = await connect(tmpDir.name);
const table = await bucketTable(conn);
@@ -5219,7 +5219,7 @@ describe("LSM merge insert", () => {
).rejects.toThrow();
});
it("auto-routes reads through the MemWAL scanner", async () => {
it("auto-routes reads through the RTE scanner", async () => {
const conn = await connect(tmpDir.name);
const table = await bucketTable(conn); // base ids "a", "b"
@@ -5233,24 +5233,22 @@ describe("LSM merge insert", () => {
const lsm = await table.query().toArray();
expect(lsm.map((r) => r.id).sort()).toEqual(["a", "b", "c"]);
// useMemWal(false) bypasses the MemWAL and reads the base table only.
const baseOnly = await table.query().useMemWal(false).toArray();
// useRTE(false) bypasses the RTE and reads the base table only.
const baseOnly = await table.query().useRTE(false).toArray();
expect(baseOnly.map((r) => r.id).sort()).toEqual(["a", "b"]);
});
it("reads the base table when no MemWAL spec is installed", async () => {
it("reads the base table when no RTE spec is installed", async () => {
const conn = await connect(tmpDir.name);
const table = await conn.createEmptyTable(
"plain",
new arrow.Schema([new arrow.Field("id", new arrow.Utf8(), false)]),
);
// No spec: default read and useMemWal(false) both succeed against the base table.
// No spec: default read and useRTE(false) both succeed against the base table.
await expect(table.query().toArray()).resolves.toBeDefined();
await expect(
table.query().useMemWal(false).toArray(),
).resolves.toBeDefined();
// useMemWal(true) demands MemWAL routing; without a spec it errors.
await expect(table.query().useMemWal(true).toArray()).rejects.toThrow();
await expect(table.query().useRTE(false).toArray()).resolves.toBeDefined();
// useRTE(true) demands RTE routing; without a spec it errors.
await expect(table.query().useRTE(true).toArray()).rejects.toThrow();
});
});
@@ -5268,42 +5266,42 @@ describe("LSM convergence and stats", () => {
new arrow.Schema([new arrow.Field("id", new arrow.Utf8(), false)]),
);
await table.setUnenforcedPrimaryKey("id");
await table.enableMemWal({ specType: "unsharded" });
await table.enableRTE({ specType: "unsharded" });
return table;
}
// These four route through the server that owns the MemWAL, so a local table
// These four route through the server that owns the RTE, so a local table
// rejects them rather than answering. What is asserted here is that the
// bindings reach the core at all; the behavior against a real endpoint is
// covered by the mocked endpoint tests in rust/lancedb/src/remote/table.rs.
it("rejects flushMemWal on a local table", async () => {
it("rejects flushRTE on a local table", async () => {
const conn = await connect(tmpDir.name);
const table = await lsmTable(conn);
await expect(table.flushMemWal()).rejects.toThrow(/not supported/i);
await expect(table.flushRTE()).rejects.toThrow(/not supported/i);
});
it("rejects compactMemWal on a local table", async () => {
it("rejects compactRTE on a local table", async () => {
const conn = await connect(tmpDir.name);
const table = await lsmTable(conn);
await expect(table.compactMemWal()).rejects.toThrow(/not supported/i);
await expect(table.compactRTE()).rejects.toThrow(/not supported/i);
});
it("rejects getMemWalStats on a local table", async () => {
it("rejects getRTEStats on a local table", async () => {
const conn = await connect(tmpDir.name);
const table = await lsmTable(conn);
await expect(table.getMemWalStats()).rejects.toThrow(/not supported/i);
await expect(table.getMemWalStats(true)).rejects.toThrow(/not supported/i);
await expect(table.getRTEStats()).rejects.toThrow(/not supported/i);
await expect(table.getRTEStats(true)).rejects.toThrow(/not supported/i);
});
it("rejects checkpointMemWal on a local table", async () => {
it("rejects checkpointRTE on a local table", async () => {
const conn = await connect(tmpDir.name);
const table = await lsmTable(conn);
// checkpointMemWal freezes first, so it surfaces flushMemWal's rejection.
await expect(table.checkpointMemWal()).rejects.toThrow(/not supported/i);
// checkpointRTE freezes first, so it surfaces flushRTE's rejection.
await expect(table.checkpointRTE()).rejects.toThrow(/not supported/i);
});
});
nodejs/lancedb/index.ts
+2 -2
View File
@@ -168,8 +168,8 @@ export {
WriteProgress,
FtsToken,
TokenizeTableOptions,
MemWalWriteSpec,
MemWalStats,
RTEWriteSpec,
RTEStats,
TableShardStats,
SsTableStats,
MemtableStats,
nodejs/lancedb/merge.ts
+8 -8
View File
@@ -88,27 +88,27 @@ export class MergeInsertBuilder {
);
}
/**
* Control MemWAL routing for this merge.
* Control RTE routing for this merge.
*
* By default (unset), a `mergeInsert` on a table with a MemWAL write spec is
* By default (unset), a `mergeInsert` on a table with an RTE write spec is
* routed through Lance's MemWAL shard writer, and a table without one uses the
* standard path.
*
* @param enable - `true` forces MemWAL routing and errors if the table has no
* MemWAL write spec. `false` forces the standard write path even when a spec is set.
* @param enable - `true` forces RTE routing and errors if the table has no
* RTE write spec. `false` forces the standard write path even when a spec is set.
*/
useMemWal(enable: boolean): MergeInsertBuilder {
return new MergeInsertBuilder(this.#native.useMemWal(enable), this.#schema);
useRTE(enable: boolean): MergeInsertBuilder {
return new MergeInsertBuilder(this.#native.useRTE(enable), this.#schema);
}
/**
* Controls how an LSM merge checks that its input targets a single shard.
*
* When a table has a MemWAL write spec, every row in a `mergeInsert` call must
* When a table has an RTE write spec, every row in a `mergeInsert` call must
* route to the same shard. When `true` (the default), every row is inspected
* to verify this. When `false`, only the first row is inspected and the
* shard it routes to is used for the whole input — a faster path for callers
* that have already pre-sharded their input. Has no effect on tables without
* a MemWAL write spec.
* an RTE write spec.
*
* @param validateSingleShard - Whether to check every row routes to one shard. Defaults to `true`.
*/
nodejs/lancedb/query.ts
+15 -15
View File
@@ -677,26 +677,26 @@ export class StandardQueryBase<
}
/**
* Control MemWAL read routing for this query.
* Control RTE read routing for this query.
*
* By default (unset), when the table carries a MemWAL write spec (see
* {@link Table#enableMemWal}), reads are routed through the LSM scanner so
* By default (unset), when the table carries an RTE write spec (see
* {@link Table#enableRTE}), reads are routed through the LSM scanner so
* they also return data written via the `mergeInsert` LSM path that has not yet
* been compacted into the base table (the active/frozen in-memory memtables and
* the SSTables), deduplicated by primary key; a table without a spec
* reads the base table.
*
* @param enable - `true` forces the LSM scanner and errors if the table has no
* MemWAL write spec. `false` bypasses the MemWAL and reads the base table only,
* RTE write spec. `false` bypasses the RTE and reads the base table only,
* even when a spec is present.
*
* Note: the LSM scanner does not support every query shape (e.g. reranking,
* hybrid search, `orderBy`). On a MemWAL table those shapes error unless
* `useMemWal(false)` is set, because a base-only read would silently exclude
* un-compacted MemWAL data.
* hybrid search, `orderBy`). On an RTE table those shapes error unless
* `useRTE(false)` is set, because a base-only read would silently exclude
* un-compacted RTE data.
*/
useMemWal(enable: boolean): this {
this.doCall((inner: NativeQueryType) => inner.useMemWal(enable));
useRTE(enable: boolean): this {
this.doCall((inner: NativeQueryType) => inner.useRTE(enable));
return this;
}
}
@@ -776,7 +776,7 @@ export class VectorQuery extends StandardQueryBase<NativeVectorQuery> {
* probing may continue when too few candidates are found.
* Any explicit effort, including 0.5, overrides nprobes in either call order.
* Explicit minimum and maximum bounds still apply. This is not a recall
* guarantee and does not change refinement or HNSW ef. The MemWAL LSM
* guarantee and does not change refinement or HNSW ef. The RTE LSM
* scanner does not yet support this option.
*
* @example
@@ -1082,16 +1082,16 @@ export class TakeQuery extends QueryBase<NativeTakeQuery> {
}
/**
* Control MemWAL read routing for this take query.
* Control RTE read routing for this take query.
*
* `false` bypasses the MemWAL and reads the base table only — the escape hatch,
* `false` bypasses the RTE and reads the base table only — the escape hatch,
* since take-by-row-id/offset is not supported on the LSM scanner and, on a
* MemWAL table, auto-routes to it and errors otherwise.
* RTE table, auto-routes to it and errors otherwise.
*
* @param enable - `false` reads the base table only.
*/
useMemWal(enable: boolean): this {
this.doCall((inner: NativeTakeQuery) => inner.useMemWal(enable));
useRTE(enable: boolean): this {
this.doCall((inner: NativeTakeQuery) => inner.useRTE(enable));
return this;
}
}
nodejs/lancedb/table.ts
+54 -54
View File
@@ -35,9 +35,9 @@ import {
FunctionErrorsOptions,
IndexConfig,
IndexStatistics,
MemWalStats,
Branches as NativeBranches,
OptimizeStats,
RTEStats,
RefreshColumnResult,
RefreshMaterializedViewResult,
TableStatistics,
@@ -61,7 +61,7 @@ export { IndexConfig } from "./native";
export {
TableShardStats,
SsTableStats,
MemWalStats,
RTEStats,
MemtableStats,
} from "./native";
@@ -207,8 +207,8 @@ export type TokenizeTableOptions =
};
/**
* Specification selecting Lance's MemWAL LSM-style write path for
* `mergeInsert`.
* Specification enabling the realtime engine (RTE), which routes
* `mergeInsert` through Lance's MemWAL LSM-style write path.
*
* `specType` is `"bucket"`, `"identity"`, or `"unsharded"`. For `"bucket"`,
* `column` and `numBuckets` are required; for `"identity"`, `column` is
@@ -217,7 +217,7 @@ export type TokenizeTableOptions =
* `column` value, or upserts of that key can land in different shards and a
* stale version can win).
*/
export interface MemWalWriteSpec {
export interface RTEWriteSpec {
/** One of `"bucket"`, `"identity"`, or `"unsharded"`. */
specType: "bucket" | "identity" | "unsharded";
/** Bucket and identity variants: the sharding column. */
@@ -225,7 +225,7 @@ export interface MemWalWriteSpec {
/** Bucket variant: the number of buckets, in `[1, 1024]`. */
numBuckets?: number;
/**
* Indexes the MemWAL keeps up to date. Omit for every index the table has,
* Indexes the RTE keeps up to date. Omit for every index the table has,
* including ones created later. Pass `[]` for none.
*/
maintainedIndexes?: string[];
@@ -758,10 +758,10 @@ export abstract class Table {
*/
abstract setUnenforcedPrimaryKey(columns: string | string[]): Promise<void>;
/**
* Install an {@link MemWalWriteSpec} on this table, selecting Lance's MemWAL
* Install an {@link RTEWriteSpec} on this table, selecting Lance's MemWAL
* LSM-style write path for future `mergeInsert` calls.
*
* `MemWalWriteSpec` chooses one of three sharding strategies via `specType`:
* `RTEWriteSpec` chooses one of three sharding strategies via `specType`:
*
* - `"bucket"` — hash-bucket writes by the single-column unenforced primary
* key (`column` and `numBuckets` required).
@@ -773,15 +773,15 @@ export abstract class Table {
* requires it to be the single column being bucketed.
*
* Omitting `maintainedIndexes` maintains every index the table has,
* including ones created later, and skips a kind the MemWAL cannot maintain.
* including ones created later, and skips a kind the RTE cannot maintain.
* Naming them pins an exact set, and a still-building index is rejected
* rather than quietly omitted.
* @param {MemWalWriteSpec} spec The sharding spec to install.
* @param {RTEWriteSpec} spec The sharding spec to install.
* @returns {Promise<void>}
* @example
* ```ts
* await table.setUnenforcedPrimaryKey("id");
* await table.enableMemWal({
* await table.enableRTE({
* specType: "bucket",
* column: "id",
* numBuckets: 16,
@@ -789,36 +789,36 @@ export abstract class Table {
* });
* ```
*/
abstract enableMemWal(spec: MemWalWriteSpec): Promise<void>;
abstract enableRTE(spec: RTEWriteSpec): Promise<void>;
/**
* Remove the {@link MemWalWriteSpec} from this table, reverting to the standard
* Remove the {@link RTEWriteSpec} from this table, reverting to the standard
* `mergeInsert` write path.
*
* Errors if no spec is currently set.
* @returns {Promise<void>}
*/
abstract disableMemWal(): Promise<void>;
abstract disableRTE(): Promise<void>;
/**
* Read the {@link MemWalWriteSpec} currently installed on this table.
* Read the {@link RTEWriteSpec} currently installed on this table.
*
* Resolves to `undefined` when the MemWAL write path is not enabled (no
* spec has been set, or it was removed with {@link Table#disableMemWal}).
* Resolves to `undefined` when the RTE write path is not enabled (no
* spec has been set, or it was removed with {@link Table#disableRTE}).
* The spec is the one installed, including its maintained-index selection:
* an absent `maintainedIndexes` for every index the table has, an empty
* array for none.
* @returns {Promise<MemWalWriteSpec | undefined>}
* @returns {Promise<RTEWriteSpec | undefined>}
*/
abstract getMemWalSpec(): Promise<MemWalWriteSpec | undefined>;
abstract getRTESpec(): Promise<RTEWriteSpec | undefined>;
/**
* Drain and close any cached MemWAL shard writers held for this table.
* Drain and close any cached RTE shard writers held for this table.
*
* When an {@link MemWalWriteSpec} is installed, `mergeInsert` opens MemWAL
* When an {@link RTEWriteSpec} is installed, `mergeInsert` opens RTE
* shard writers and caches them for reuse across calls. This closes them,
* flushing pending data; writers reopen lazily on the next `mergeInsert`.
* It is a no-op when no writers are cached.
* @returns {Promise<void>}
*/
abstract closeMemWalWriters(): Promise<void>;
abstract closeRTEWriters(): Promise<void>;
/**
* Freeze every table shard's active memtable into a new SSTable.
*
@@ -826,18 +826,18 @@ export abstract class Table {
* so this is safe to call repeatedly.
* @returns {Promise<void>}
*/
abstract flushMemWal(): Promise<void>;
abstract flushRTE(): Promise<void>;
/**
* Trigger a background SSTable compaction pass per table shard.
*
* Returns once the passes are *dispatched*, not once they finish — watch
* {@link Table#getMemWalStats} for progress, or use
* {@link Table#checkpointMemWal} to wait for convergence.
* {@link Table#getRTEStats} for progress, or use
* {@link Table#checkpointRTE} to wait for convergence.
* @returns {Promise<void>}
*/
abstract compactMemWal(): Promise<void>;
abstract compactRTE(): Promise<void>;
/**
* Converge this table's MemWAL write path into its base table.
* Converge this table's RTE write path into its base table.
*
* Freezes once, then triggers compaction and polls until the SSTables that existed
* at the start is gone. The target set is fixed at the start, so
@@ -852,26 +852,26 @@ export abstract class Table {
* @returns {Promise<void>}
* @example
* ```ts
* const before = await table.getMemWalStats();
* await table.checkpointMemWal();
* const after = await table.getMemWalStats();
* const before = await table.getRTEStats();
* await table.checkpointRTE();
* const after = await table.getRTEStats();
* ```
*/
abstract checkpointMemWal(): Promise<void>;
abstract checkpointRTE(): Promise<void>;
/**
* Read live per-table-shard MemWAL state.
* Read live per-table-shard RTE state.
*
* Answers "how far behind is my fresh tier", "which table shard is hot", and
* "why is my fresh-tier vector search brute-force". Mutates no table state.
*
* Resolves to `undefined` only when the MemWAL write path is not enabled.
* Resolves to `undefined` only when the RTE write path is not enabled.
* @param {boolean} includeSstableRows Also count rows per SSTable.
* Off by default because each count opens an uncached Lance dataset.
* @returns {Promise<MemWalStats | undefined>}
* @returns {Promise<RTEStats | undefined>}
*/
abstract getMemWalStats(
abstract getRTEStats(
includeSstableRows?: boolean,
): Promise<MemWalStats | undefined>;
): Promise<RTEStats | undefined>;
/** Retrieve the version of the table */
abstract version(): Promise<number>;
@@ -1480,43 +1480,43 @@ export class LocalTable extends Table {
return await this.inner.setUnenforcedPrimaryKey(cols);
}
async enableMemWal(spec: MemWalWriteSpec): Promise<void> {
return await this.inner.enableMemWal(spec);
async enableRTE(spec: RTEWriteSpec): Promise<void> {
return await this.inner.enableRTE(spec);
}
async disableMemWal(): Promise<void> {
return await this.inner.disableMemWal();
async disableRTE(): Promise<void> {
return await this.inner.disableRTE();
}
async getMemWalSpec(): Promise<MemWalWriteSpec | undefined> {
async getRTESpec(): Promise<RTEWriteSpec | undefined> {
// The native binding types `specType` as a plain `string`; narrow it back
// to the public union. The Rust `From` impl only ever emits one of the
// three valid values, so the cast is safe.
return ((await this.inner.getMemWalSpec()) ?? undefined) as
| MemWalWriteSpec
return ((await this.inner.getRTESpec()) ?? undefined) as
| RTEWriteSpec
| undefined;
}
async closeMemWalWriters(): Promise<void> {
return await this.inner.closeMemWalWriters();
async closeRTEWriters(): Promise<void> {
return await this.inner.closeRTEWriters();
}
async flushMemWal(): Promise<void> {
return await this.inner.flushMemWal();
async flushRTE(): Promise<void> {
return await this.inner.flushRTE();
}
async compactMemWal(): Promise<void> {
return await this.inner.compactMemWal();
async compactRTE(): Promise<void> {
return await this.inner.compactRTE();
}
async checkpointMemWal(): Promise<void> {
return await this.inner.checkpointMemWal();
async checkpointRTE(): Promise<void> {
return await this.inner.checkpointRTE();
}
async getMemWalStats(
async getRTEStats(
includeSstableRows: boolean = false,
): Promise<MemWalStats | undefined> {
return (await this.inner.getMemWalStats(includeSstableRows)) ?? undefined;
): Promise<RTEStats | undefined> {
return (await this.inner.getRTEStats(includeSstableRows)) ?? undefined;
}
async version(): Promise<number> {
nodejs/src/merge.rs
+3 -3
View File
@@ -50,10 +50,10 @@ impl NativeMergeInsertBuilder {
this
}
#[napi(js_name = "useMemWal")]
pub fn use_memwal(&self, enable: bool) -> Self {
#[napi(js_name = "useRTE")]
pub fn use_rte(&self, enable: bool) -> Self {
let mut this = self.clone();
this.inner.use_memwal(enable);
this.inner.use_rte(enable);
this
}
nodejs/src/query.rs
+9 -9
View File
@@ -168,9 +168,9 @@ impl Query {
self.inner = self.inner.clone().with_row_id();
}
#[napi(js_name = "useMemWal")]
pub fn use_memwal(&mut self, enable: bool) {
self.inner = self.inner.clone().use_memwal(enable);
#[napi(js_name = "useRTE")]
pub fn use_rte(&mut self, enable: bool) {
self.inner = self.inner.clone().use_rte(enable);
}
#[napi]
@@ -399,9 +399,9 @@ impl VectorQuery {
self.inner = self.inner.clone().with_row_id();
}
#[napi(js_name = "useMemWal")]
pub fn use_memwal(&mut self, enable: bool) {
self.inner = self.inner.clone().use_memwal(enable);
#[napi(js_name = "useRTE")]
pub fn use_rte(&mut self, enable: bool) {
self.inner = self.inner.clone().use_rte(enable);
}
#[napi]
@@ -514,9 +514,9 @@ impl TakeQuery {
self.inner = self.inner.clone().with_row_id();
}
#[napi(js_name = "useMemWal")]
pub fn use_memwal(&mut self, enable: bool) {
self.inner = self.inner.clone().use_memwal(enable);
#[napi(js_name = "useRTE")]
pub fn use_rte(&mut self, enable: bool) {
self.inner = self.inner.clone().use_rte(enable);
}
#[napi(catch_unwind)]
nodejs/src/table.rs
+47 -50
View File
@@ -580,60 +580,57 @@ impl Table {
.default_error()
}
#[napi(catch_unwind, js_name = "enableMemWal")]
pub async fn enable_memwal(&self, spec: MemWalWriteSpec) -> napi::Result<()> {
let native_spec = lancedb::table::MemWalWriteSpec::try_from(spec)?;
#[napi(catch_unwind, js_name = "enableRTE")]
pub async fn enable_rte(&self, spec: RTEWriteSpec) -> napi::Result<()> {
let native_spec = lancedb::table::RTEWriteSpec::try_from(spec)?;
self.inner_ref()?
.enable_memwal(native_spec)
.enable_rte(native_spec)
.await
.default_error()
}
#[napi(catch_unwind, js_name = "disableMemWal")]
pub async fn disable_memwal(&self) -> napi::Result<()> {
self.inner_ref()?.disable_memwal().await.default_error()
#[napi(catch_unwind, js_name = "disableRTE")]
pub async fn disable_rte(&self) -> napi::Result<()> {
self.inner_ref()?.disable_rte().await.default_error()
}
#[napi(catch_unwind, js_name = "getMemWalSpec")]
pub async fn get_memwal_spec(&self) -> napi::Result<Option<MemWalWriteSpec>> {
let spec = self.inner_ref()?.get_memwal_spec().await.default_error()?;
Ok(spec.map(MemWalWriteSpec::from))
#[napi(catch_unwind, js_name = "getRTESpec")]
pub async fn get_rte_spec(&self) -> napi::Result<Option<RTEWriteSpec>> {
let spec = self.inner_ref()?.get_rte_spec().await.default_error()?;
Ok(spec.map(RTEWriteSpec::from))
}
#[napi(catch_unwind, js_name = "closeMemWalWriters")]
pub async fn close_memwal_writers(&self) -> napi::Result<()> {
self.inner_ref()?
.close_memwal_writers()
.await
.default_error()
#[napi(catch_unwind, js_name = "closeRTEWriters")]
pub async fn close_rte_writers(&self) -> napi::Result<()> {
self.inner_ref()?.close_rte_writers().await.default_error()
}
#[napi(catch_unwind, js_name = "flushMemWal")]
pub async fn flush_memwal(&self) -> napi::Result<()> {
self.inner_ref()?.flush_memwal().await.default_error()
#[napi(catch_unwind, js_name = "flushRTE")]
pub async fn flush_rte(&self) -> napi::Result<()> {
self.inner_ref()?.flush_rte().await.default_error()
}
#[napi(catch_unwind, js_name = "compactMemWal")]
pub async fn compact_memwal(&self) -> napi::Result<()> {
self.inner_ref()?.compact_memwal().await.default_error()
#[napi(catch_unwind, js_name = "compactRTE")]
pub async fn compact_rte(&self) -> napi::Result<()> {
self.inner_ref()?.compact_rte().await.default_error()
}
#[napi(catch_unwind, js_name = "checkpointMemWal")]
pub async fn checkpoint_memwal(&self) -> napi::Result<()> {
self.inner_ref()?.checkpoint_memwal().await.default_error()
#[napi(catch_unwind, js_name = "checkpointRTE")]
pub async fn checkpoint_rte(&self) -> napi::Result<()> {
self.inner_ref()?.checkpoint_rte().await.default_error()
}
#[napi(catch_unwind, js_name = "getMemWalStats")]
pub async fn get_memwal_stats(
#[napi(catch_unwind, js_name = "getRTEStats")]
pub async fn get_rte_stats(
&self,
include_sstable_rows: bool,
) -> napi::Result<Option<MemWalStats>> {
) -> napi::Result<Option<RTEStats>> {
let stats = self
.inner_ref()?
.get_memwal_stats(include_sstable_rows)
.get_rte_stats(include_sstable_rows)
.await
.default_error()?;
Ok(stats.map(MemWalStats::from))
Ok(stats.map(RTEStats::from))
}
#[napi(catch_unwind)]
@@ -946,53 +943,53 @@ impl From<LanceDbFtsToken> for FtsToken {
}
}
/// Specification selecting Lance's MemWAL LSM-style write path for
/// `mergeInsert`.
/// Specification enabling the realtime engine (RTE), which routes
/// `mergeInsert` through Lance's MemWAL LSM-style write path.
///
/// `specType` must be `"bucket"`, `"identity"`, or `"unsharded"`. For
/// `"bucket"`, `column` and `numBuckets` are required; for `"identity"`,
/// `column` is required.
#[napi(object)]
#[napi(object, js_name = "RTEWriteSpec")]
#[derive(Clone, Debug)]
pub struct MemWalWriteSpec {
pub struct RTEWriteSpec {
/// One of `"bucket"`, `"identity"`, or `"unsharded"`.
pub spec_type: String,
/// Bucket and identity variants: the sharding column.
pub column: Option<String>,
/// Bucket variant: the number of buckets, in `[1, 1024]`.
pub num_buckets: Option<u32>,
/// Indexes the MemWAL keeps up to date. Omitted resolves every
/// Indexes the RTE keeps up to date. Omitted resolves every
/// maintainable index on install; an empty array means none.
pub maintained_indexes: Option<Vec<String>>,
/// Default `ShardWriter` configuration recorded in the MemWAL index.
pub writer_config_defaults: Option<HashMap<String, String>>,
}
impl TryFrom<MemWalWriteSpec> for lancedb::table::MemWalWriteSpec {
impl TryFrom<RTEWriteSpec> for lancedb::table::RTEWriteSpec {
type Error = napi::Error;
fn try_from(value: MemWalWriteSpec) -> napi::Result<Self> {
fn try_from(value: RTEWriteSpec) -> napi::Result<Self> {
let writer_config_defaults = value.writer_config_defaults.unwrap_or_default();
let spec = match value.spec_type.as_str() {
"bucket" => {
let column = value.column.ok_or_else(|| {
napi::Error::from_reason("MemWalWriteSpec bucket requires `column`")
napi::Error::from_reason("RTEWriteSpec bucket requires `column`")
})?;
let num_buckets = value.num_buckets.ok_or_else(|| {
napi::Error::from_reason("MemWalWriteSpec bucket requires `numBuckets`")
napi::Error::from_reason("RTEWriteSpec bucket requires `numBuckets`")
})?;
Self::bucket(column, num_buckets)
}
"identity" => {
let column = value.column.ok_or_else(|| {
napi::Error::from_reason("MemWalWriteSpec identity requires `column`")
napi::Error::from_reason("RTEWriteSpec identity requires `column`")
})?;
Self::identity(column)
}
"unsharded" => Self::unsharded(),
other => {
return Err(napi::Error::from_reason(format!(
"MemWalWriteSpec `specType` must be 'bucket', 'identity', or 'unsharded', got '{}'",
"RTEWriteSpec `specType` must be 'bucket', 'identity', or 'unsharded', got '{}'",
other
)));
}
@@ -1003,9 +1000,9 @@ impl TryFrom<MemWalWriteSpec> for lancedb::table::MemWalWriteSpec {
}
}
impl From<lancedb::table::MemWalWriteSpec> for MemWalWriteSpec {
fn from(spec: lancedb::table::MemWalWriteSpec) -> Self {
use lancedb::table::MemWalWriteSpec as Native;
impl From<lancedb::table::RTEWriteSpec> for RTEWriteSpec {
fn from(spec: lancedb::table::RTEWriteSpec) -> Self {
use lancedb::table::RTEWriteSpec as Native;
match spec {
Native::Bucket {
column,
@@ -1148,19 +1145,19 @@ impl From<lancedb::table::TableShardStats> for TableShardStats {
}
}
/// Live per-table-shard MemWAL state, as returned by `Table#getMemWalStats`.
/// Live per-table-shard RTE state, as returned by `Table#getRTEStats`.
///
/// Nothing here is derived: sums and differences (total SSTable bytes, WAL lag) are
/// the caller's to compute.
#[napi(object)]
#[napi(object, js_name = "RTEStats")]
#[derive(Clone, Debug)]
pub struct MemWalStats {
pub struct RTEStats {
/// One entry per table shard backing this table.
pub table_shards: Vec<TableShardStats>,
}
impl From<lancedb::table::MemWalStats> for MemWalStats {
fn from(stats: lancedb::table::MemWalStats) -> Self {
impl From<lancedb::table::RTEStats> for RTEStats {
fn from(stats: lancedb::table::RTEStats) -> Self {
Self {
table_shards: stats.table_shards.into_iter().map(Into::into).collect(),
}
python/python/lancedb/__init__.py
+2 -2
View File
@@ -15,7 +15,7 @@ from ._lancedb import FtsToken
from ._lancedb import FunctionErrorFragment as FunctionErrorFragment
from ._lancedb import FunctionErrorRecord as FunctionErrorRecord
from ._lancedb import FunctionErrors as FunctionErrors
from ._lancedb import MemWalWriteSpec
from ._lancedb import RTEWriteSpec
from ._lancedb import tokenize as _tokenize
from .common import URI, sanitize_uri
from urllib.parse import urlparse
@@ -610,7 +610,7 @@ __all__ = [
"SqlQuery",
"LanceDBConnection",
"LanceNamespaceDBConnection",
"MemWalWriteSpec",
"RTEWriteSpec",
"RemoteDBConnection",
"Session",
"Table",
python/python/lancedb/_lancedb.pyi
+24 -28
View File
@@ -538,17 +538,17 @@ class Table:
async def initial_storage_options(self) -> Optional[Dict[str, str]]: ...
async def latest_storage_options(self) -> Optional[Dict[str, str]]: ...
async def set_unenforced_primary_key(self, columns: List[str]) -> None: ...
async def enable_memwal(self, spec: MemWalWriteSpec) -> None: ...
async def disable_memwal(self) -> None: ...
async def get_memwal_spec(self) -> Optional[MemWalWriteSpec]: ...
async def enable_rte(self, spec: RTEWriteSpec) -> None: ...
async def disable_rte(self) -> None: ...
async def get_rte_spec(self) -> Optional[RTEWriteSpec]: ...
# WAL-PK-FUSION: delete both.
def hybrid_pk_fusion_learned(self) -> bool: ...
def note_hybrid_pk_fusion(self) -> None: ...
async def checkpoint_memwal(self) -> None: ...
async def flush_memwal(self) -> None: ...
async def compact_memwal(self) -> None: ...
async def get_memwal_stats(self, include_sstable_rows: bool) -> Optional[dict]: ...
async def close_memwal_writers(self) -> None: ...
async def checkpoint_rte(self) -> None: ...
async def flush_rte(self) -> None: ...
async def compact_rte(self) -> None: ...
async def get_rte_stats(self, include_sstable_rows: bool) -> Optional[dict]: ...
async def close_rte_writers(self) -> None: ...
@property
def tags(self) -> Tags: ...
@property
@@ -661,7 +661,7 @@ class Query:
def fast_search(self): ...
def with_row_id(self): ...
def postfilter(self): ...
def use_memwal(self, enable: bool): ...
def use_rte(self, enable: bool): ...
def nearest_to(self, query_vec: pa.Array) -> VectorQuery: ...
def nearest_to_text(self, query: dict) -> FTSQuery: ...
def order_by(self, ordering: Optional[List[ColumnOrdering]]): ...
@@ -678,7 +678,7 @@ class Query:
class TakeQuery:
def select(self, columns: List[str]): ...
def with_row_id(self): ...
def use_memwal(self, enable: bool): ...
def use_rte(self, enable: bool): ...
async def output_schema(self) -> pa.Schema: ...
async def execute(self) -> RecordBatchStream: ...
async def explain_plan(self, verbose: Optional[bool]) -> str: ...
@@ -697,7 +697,7 @@ class FTSQuery:
def fast_search(self): ...
def with_row_id(self): ...
def postfilter(self): ...
def use_memwal(self, enable: bool): ...
def use_rte(self, enable: bool): ...
def get_query(self) -> str: ...
def add_query_vector(self, query_vec: pa.Array) -> None: ...
def nearest_to(self, query_vec: pa.Array) -> HybridQuery: ...
@@ -725,7 +725,7 @@ class VectorQuery:
def column(self, column: str): ...
def distance_type(self, distance_type: str): ...
def postfilter(self): ...
def use_memwal(self, enable: bool): ...
def use_rte(self, enable: bool): ...
def refine_factor(self, refine_factor: int): ...
def nprobes(self, nprobes: int): ...
def search_effort(self, search_effort: Optional[float]): ...
@@ -750,7 +750,7 @@ class HybridQuery:
def fast_search(self): ...
def with_row_id(self): ...
def postfilter(self): ...
def use_memwal(self, enable: bool): ...
def use_rte(self, enable: bool): ...
def distance_type(self, distance_type: str): ...
def refine_factor(self, refine_factor: int): ...
def nprobes(self, nprobes: int): ...
@@ -778,7 +778,7 @@ class PyQueryRequest:
select_source_columns: Optional[Dict[str, str]]
fast_search: Optional[bool]
with_row_id: Optional[bool]
use_memwal: Optional[bool]
use_rte: Optional[bool]
column: Optional[str]
query_vector: Optional[List[pa.Array]]
nprobes: Optional[int]
@@ -839,26 +839,22 @@ class FtsToken:
text: str
position: int
class MemWalWriteSpec:
"""Specification selecting Lance's MemWAL LSM-style write path for
`merge_insert`."""
class RTEWriteSpec:
"""Specification enabling the realtime engine (RTE), which routes
`merge_insert` through Lance's MemWAL LSM-style write path."""
@staticmethod
def bucket(column: str, num_buckets: int) -> "MemWalWriteSpec": ...
def bucket(column: str, num_buckets: int) -> "RTEWriteSpec": ...
@staticmethod
def identity(column: str) -> "MemWalWriteSpec": ...
def identity(column: str) -> "RTEWriteSpec": ...
@staticmethod
def unsharded() -> "MemWalWriteSpec": ...
def with_maintained_indexes(
self, indexes: Optional[List[str]]
) -> "MemWalWriteSpec":
"""Set which indexes the MemWAL keeps up to date. None resolves every
def unsharded() -> "RTEWriteSpec": ...
def with_maintained_indexes(self, indexes: Optional[List[str]]) -> "RTEWriteSpec":
"""Set which indexes the RTE keeps up to date. None resolves every
index on the table at install, failing if one cannot be maintained;
a list is verbatim, empty means none."""
...
def with_writer_config_defaults(
self, defaults: Dict[str, str]
) -> "MemWalWriteSpec":
def with_writer_config_defaults(self, defaults: Dict[str, str]) -> "RTEWriteSpec":
"""Return a copy of this spec recording the given default
`ShardWriter` configuration in the MemWAL index."""
...
@@ -872,7 +868,7 @@ class MemWalWriteSpec:
def num_buckets(self) -> Optional[int]: ...
@property
def maintained_indexes(self) -> Optional[List[str]]:
"""Indexes the MemWAL keeps up to date, or None for every supported one."""
"""Indexes the RTE keeps up to date, or None for every supported one."""
...
@property
def writer_config_defaults(self) -> Dict[str, str]: ...
python/python/lancedb/_wal_hybrid.py
+19 -19
View File
@@ -1,14 +1,14 @@
# SPDX-License-Identifier: Apache-2.0
# SPDX-FileCopyrightText: Copyright The LanceDB Authors
"""WAL-PK-FUSION: hybrid search's primary-key fallback for MemWAL tables.
"""WAL-PK-FUSION: hybrid search's primary-key fallback for RTE tables.
Hybrid search joins its vector and full-text legs on `_rowid`, which a MemWAL
Hybrid search joins its vector and full-text legs on `_rowid`, which an RTE
table cannot supply: the fresh tier has no stable row id, so the server rejects
`with_row_id` before it plans. When that happens the legs are joined on a
surrogate `_rowid` built from the table's primary key instead.
This is a stopgap until MemWAL supports `_rowid`, and everything it needs lives
This is a stopgap until RTE supports `_rowid`, and everything it needs lives
here so it can be removed in one pass:
1. Delete this file and `tests/test_wal_hybrid.py`.
@@ -69,21 +69,21 @@ def _pk_columns(schema: pa.Schema) -> List[str]:
# builds it from the same words -- so this substring survives either path and
# the wrapping each adds. It is a coupling to a message rather than a code,
# because the refusal arrives as a generic 400; if the server ever rewords it
# the fallback stops firing and hybrid on a MemWAL table errors visibly again,
# the fallback stops firing and hybrid on an RTE table errors visibly again,
# which is loud rather than silently wrong.
_WAL_ROW_ID_REFUSAL = "does not support with_row_id"
WAL_ROW_ID_UNSUPPORTED = (
"hybrid search on a MemWAL table cannot return _rowid: the fresh tier has "
"hybrid search on an RTE table cannot return _rowid: the fresh tier has "
"no stable row id, and the ids the fusion joins on are synthesized from "
"the primary key. Set use_memwal(False) to read the base table only (results "
"will exclude un-compacted MemWAL data)"
"the primary key. Set use_rte(False) to read the base table only (results "
"will exclude un-compacted RTE data)"
)
def _is_wal_row_id_refusal(error: BaseException) -> bool:
"""Whether `error` is the server declining `_rowid` because the table is
MemWAL-backed, as opposed to any other failure the query might hit."""
RTE-backed, as opposed to any other failure the query might hit."""
return _WAL_ROW_ID_REFUSAL in str(error)
@@ -112,7 +112,7 @@ def _stamp_surrogate_row_ids(
"""Give each table a `_rowid` column derived from its primary key.
The fusion only ever compares that column for equality — to deduplicate, to
group by in a reranker, and to restore the original scores — so on a MemWAL
group by in a reranker, and to restore the original scores — so on an RTE
table, where the server has no stable row id to hand out, any injective
mapping of the primary key does the same job. Ids are assigned in first-seen
order across `tables`, so passing them in fusion order (vector, then FTS)
@@ -124,7 +124,7 @@ def _stamp_surrogate_row_ids(
"""
if not pk_columns:
raise ValueError(
"hybrid search on a MemWAL table needs an unenforced primary key to "
"hybrid search on an RTE table needs an unenforced primary key to "
"deduplicate on, and this table declares none"
)
@@ -151,13 +151,13 @@ def _stamp_surrogate_row_ids(
return stamped
def pk_fusion_learned(table, use_memwal: Optional[bool]) -> bool:
def pk_fusion_learned(table, use_rte: Optional[bool]) -> bool:
"""Whether a previous hybrid query on `table` was refused `_rowid`, so this
one should go straight to the primary key.
Learned from a refusal, never probed, so asking is free.
"""
if table is None or use_memwal is False:
if table is None or use_rte is False:
return False
# Only an unambiguous True forks to the key: a table type that answers the
# question loosely keeps the `_rowid` path.
@@ -165,7 +165,7 @@ def pk_fusion_learned(table, use_memwal: Optional[bool]) -> bool:
class PkFusion:
"""How the legs of a hybrid query on a MemWAL table are joined: on a
"""How the legs of a hybrid query on an RTE table are joined: on a
surrogate `_rowid` built from the primary key."""
def __init__(self, pk_columns: List[str]):
@@ -190,16 +190,16 @@ class PkFusion:
# has turned `_with_row_id` on — including `return_score="all"`.
if blob_v2_projection_sources(schema, projection):
raise NotImplementedError(
"hybrid search cannot project a blob column on a MemWAL table: "
"hybrid search cannot project a blob column on an RTE table: "
"fetching blobs needs a real _rowid, and the fresh tier has no "
"stable one. Set use_memwal(False) to read the base table only "
"(results will exclude un-compacted MemWAL data)"
"stable one. Set use_rte(False) to read the base table only "
"(results will exclude un-compacted RTE data)"
)
return cls(_pk_columns(schema))
def inject(self, columns):
"""`columns` with the key columns it is missing, to project in place of
the `_rowid` a MemWAL table would refuse."""
the `_rowid` an RTE table would refuse."""
columns, self.injected = _with_pk_columns(columns, self.pk_columns)
return columns
@@ -226,7 +226,7 @@ class PkFusion:
def _check_refusal(error: Exception, *, fused: bool, caller_requested_row_id: bool):
"""Re-raise `error` unless it is a MemWAL `_rowid` refusal worth retrying on
"""Re-raise `error` unless it is an RTE `_rowid` refusal worth retrying on
the primary key."""
if fused or not _is_wal_row_id_refusal(error):
raise error
@@ -248,7 +248,7 @@ def with_pk_fallback(
`_rowid`.
`_rowid` is asked for first rather than paying a round trip up front to
find out: only a MemWAL table refuses, and the refusal has to be handled
find out: only an RTE table refuses, and the refusal has to be handled
anyway -- a write spec installed elsewhere can arrive between any two
queries. It costs nothing server-side: it is raised before the query is
planned. `fused` reports whether the attempt already used the key.
python/python/lancedb/merge.py
+8 -8
View File
@@ -37,7 +37,7 @@ class LanceMergeInsertBuilder(object):
self._when_not_matched_by_source_condition_expr = None
self._timeout = None
self._use_index = True
self._use_memwal = None
self._use_rte = None
self._validate_single_shard = None
def when_matched_update_all(
@@ -115,22 +115,22 @@ class LanceMergeInsertBuilder(object):
self._use_index = use_index
return self
def use_memwal(self, enable: bool) -> LanceMergeInsertBuilder:
def use_rte(self, enable: bool) -> LanceMergeInsertBuilder:
"""
Control MemWAL routing for this merge.
Control RTE routing for this merge.
By default (unset), a `merge_insert` on a table with a MemWAL write spec is
By default (unset), a `merge_insert` on a table with an RTE write spec is
routed through Lance's MemWAL shard writer, and a table without one uses
the standard path.
Parameters
----------
enable: bool
``True`` forces MemWAL routing and errors if the table has no MemWAL
``True`` forces RTE routing and errors if the table has no RTE
write spec. ``False`` forces the standard write path even when a spec
is set.
"""
self._use_memwal = enable
self._use_rte = enable
return self
def validate_single_shard(
@@ -139,13 +139,13 @@ class LanceMergeInsertBuilder(object):
"""
Controls how an LSM merge checks that its input targets a single shard.
When a table has a MemWAL write spec, every row in a `merge_insert` call
When a table has an RTE write spec, every row in a `merge_insert` call
must route to the same shard. When `True` (the default), every row is
inspected to verify this. When `False`, only the first row is inspected
and the shard it routes to is used for the whole input — a faster path
for callers that have already pre-sharded their input.
Has no effect on tables without a MemWAL write spec.
Has no effect on tables without an RTE write spec.
Parameters
----------
python/python/lancedb/permutation.py
+1 -1
View File
@@ -48,7 +48,7 @@ class PermutationBuilder:
By default, the permutation builder will create a single split that contains all
rows in the same order as the base table.
Tables with a MemWAL write spec are rejected: unflushed rows have no row id.
Tables with an RTE write spec are rejected: unflushed rows have no row id.
"""
self._async = async_permutation_builder(table)
python/python/lancedb/query.py
+39 -39
View File
@@ -887,10 +887,10 @@ class Query(pydantic.BaseModel):
# if true, will only search the indexed data
fast_search: Optional[bool] = None
# MemWAL LSM read routing: None auto-routes when the table carries a write
# RTE LSM read routing: None auto-routes when the table carries a write
# spec, True forces the LSM scanner (errors without a spec), False reads the
# base table only
use_memwal: Optional[bool] = None
use_rte: Optional[bool] = None
# size of the nearest neighbor list maintained during HNSW search
ef: Optional[int] = None
@@ -910,9 +910,9 @@ class Query(pydantic.BaseModel):
query.full_text_query = req.full_text_search
query.columns = req.select
query.with_row_id = req.with_row_id
# use_memwal is a genuine tri-state (None / True / False); preserve it as-is
# use_rte is a genuine tri-state (None / True / False); preserve it as-is
# so a round-tripped query keeps an explicit False.
query.use_memwal = req.use_memwal
query.use_rte = req.use_rte
query.vector_column = req.column
query.vector = req.query_vector
query.distance_type = req.distance_type
@@ -1085,7 +1085,7 @@ class LanceQueryBuilder(ABC):
self._with_row_address = None
self._fragments = None
self._fragment_ids = None
self._use_memwal = None
self._use_rte = None
self._vector = None
self._text = None
self._ef = None
@@ -1520,10 +1520,10 @@ class LanceQueryBuilder(ABC):
self._fragment_ids = fragment_ids
return self
def use_memwal(self, enable: bool) -> Self:
"""Control MemWAL LSM read routing for this query.
def use_rte(self, enable: bool) -> Self:
"""Control RTE LSM read routing for this query.
By default (unset), a query against a table with a MemWAL write spec is
By default (unset), a query against a table with an RTE write spec is
routed through the LSM scanner so it also returns data written via the
``merge_insert`` LSM path that has not yet been compacted into the base
table (active/frozen memtables + SSTables); a table without a
@@ -1532,8 +1532,8 @@ class LanceQueryBuilder(ABC):
Parameters
----------
enable : bool
``True`` forces the LSM scanner and errors if the table has no MemWAL
write spec. ``False`` bypasses the MemWAL and reads the base table
``True`` forces the LSM scanner and errors if the table has no RTE
write spec. ``False`` bypasses the RTE and reads the base table
only, even when a spec is present.
Returns
@@ -1541,7 +1541,7 @@ class LanceQueryBuilder(ABC):
LanceQueryBuilder
The LanceQueryBuilder object.
"""
self._use_memwal = enable
self._use_rte = enable
return self
def explain_plan(self, verbose: Optional[bool] = False) -> str:
@@ -1878,7 +1878,7 @@ class LanceVectorQueryBuilder(LanceQueryBuilder):
probing may continue when too few candidates are found. Any explicit
effort, including 0.5, overrides nprobes in either call order. Explicit
minimum and maximum bounds still apply. This is not a recall guarantee
and does not change refinement or HNSW ef. MemWAL LSM queries do not
and does not change refinement or HNSW ef. RTE LSM queries do not
yet support this option.
Examples
@@ -2055,7 +2055,7 @@ class LanceVectorQueryBuilder(LanceQueryBuilder):
with_row_address=self._with_row_address,
fragments=self._fragments,
fragment_ids=self._fragment_ids,
use_memwal=self._use_memwal,
use_rte=self._use_rte,
offset=self._offset,
fast_search=self._fast_search,
ef=self._ef,
@@ -2280,7 +2280,7 @@ class LanceFtsQueryBuilder(LanceQueryBuilder):
with_row_address=self._with_row_address,
fragments=self._fragments,
fragment_ids=self._fragment_ids,
use_memwal=self._use_memwal,
use_rte=self._use_rte,
full_text_query=FullTextSearchQuery(
query=self._query_with_phrase_semantics(), columns=self._fts_columns
),
@@ -2362,7 +2362,7 @@ class LanceEmptyQueryBuilder(LanceQueryBuilder):
with_row_address=self._with_row_address,
fragments=self._fragments,
fragment_ids=self._fragment_ids,
use_memwal=self._use_memwal,
use_rte=self._use_rte,
offset=self._offset,
order_by=self._order_by,
)
@@ -2458,7 +2458,7 @@ class LanceHybridQueryBuilder(LanceQueryBuilder):
"""Set whether to return row ids.
Recorded separately from `_with_row_id`, which `rerank(return_score=
"all")` also sets for its own use — on a MemWAL table only a caller who
"all")` also sets for its own use — on an RTE table only a caller who
asked is refused, and the reranker still falls back to the primary key.
"""
self._caller_requested_row_id = with_row_id
@@ -2728,7 +2728,7 @@ class LanceHybridQueryBuilder(LanceQueryBuilder):
if reranker.score == "all":
# WAL-PK-FUSION: without the fallback, `self.with_row_id(True)`.
# Not through `with_row_id`, which records caller intent: the
# reranker's need for row ids must not refuse a MemWAL table.
# reranker's need for row ids must not refuse an RTE table.
self._with_row_id = True
return self
@@ -2766,7 +2766,7 @@ class LanceHybridQueryBuilder(LanceQueryBuilder):
probing may continue when too few candidates are found. Any explicit
effort, including 0.5, overrides nprobes in either call order. Explicit
minimum and maximum bounds still apply. This is not a recall guarantee
and does not change refinement or HNSW ef. MemWAL LSM queries do not
and does not change refinement or HNSW ef. RTE LSM queries do not
yet support this option.
Examples
@@ -3040,7 +3040,7 @@ class LanceHybridQueryBuilder(LanceQueryBuilder):
# WAL-PK-FUSION: without the fallback, select `self._columns` as is.
self._pk_fusion = None
columns = self._columns
if _wal_hybrid.pk_fusion_learned(self._table, self._use_memwal):
if _wal_hybrid.pk_fusion_learned(self._table, self._use_rte):
self._pk_fusion = _wal_hybrid.PkFusion.for_query(
self._table.schema,
self._columns,
@@ -3057,9 +3057,9 @@ class LanceHybridQueryBuilder(LanceQueryBuilder):
if self._with_row_id and self._pk_fusion is None:
self._vector_query.with_row_id(True)
self._fts_query.with_row_id(True)
if self._use_memwal is not None:
self._vector_query.use_memwal(self._use_memwal)
self._fts_query.use_memwal(self._use_memwal)
if self._use_rte is not None:
self._vector_query.use_rte(self._use_rte)
self._fts_query.use_rte(self._use_rte)
if self._phrase_query:
self._fts_query.phrase_query(True)
if self._distance_type:
@@ -3723,12 +3723,12 @@ class AsyncStandardQuery(AsyncQueryBase):
self._inner.fast_search()
return self
def use_memwal(self, enable: bool) -> Self:
def use_rte(self, enable: bool) -> Self:
"""
Control MemWAL LSM read routing for this query.
Control RTE LSM read routing for this query.
By default (unset), a query against a table with a MemWAL write spec (see
[AsyncTable.enable_memwal][lancedb.table.AsyncTable.enable_memwal])
By default (unset), a query against a table with an RTE write spec (see
[AsyncTable.enable_rte][lancedb.table.AsyncTable.enable_rte])
is routed through the LSM scanner so it also returns data written via the
``merge_insert`` LSM path that has not yet been compacted into the base
table (the active/frozen in-memory memtables and the SSTables),
@@ -3737,11 +3737,11 @@ class AsyncStandardQuery(AsyncQueryBase):
Parameters
----------
enable : bool
``True`` forces the LSM scanner and errors if the table has no MemWAL
write spec. ``False`` bypasses the MemWAL and reads the base table
``True`` forces the LSM scanner and errors if the table has no RTE
write spec. ``False`` bypasses the RTE and reads the base table
only, even when a spec is present.
"""
self._inner.use_memwal(enable)
self._inner.use_rte(enable)
return self
def postfilter(self) -> Self:
@@ -4064,7 +4064,7 @@ class AsyncVectorQueryBase:
minimum and maximum bounds remain effective. Zero starts at the caller
minimum and one searches all partitions within the maximum. Later
probing may continue to obtain enough candidates. This is not a recall
guarantee and does not change refinement or HNSW ef. MemWAL LSM queries
guarantee and does not change refinement or HNSW ef. RTE LSM queries
do not yet support this option.
Examples
@@ -4388,7 +4388,7 @@ class AsyncHybridQuery(AsyncStandardQuery, AsyncVectorQueryBase):
# WAL-PK-FUSION: without the fallback, keep only the two `with_row_id`
# calls.
self._pk_fusion = None
if _wal_hybrid.pk_fusion_learned(self._table, fts_req.use_memwal):
if _wal_hybrid.pk_fusion_learned(self._table, fts_req.use_rte):
self._pk_fusion = _wal_hybrid.PkFusion.for_query(
await self._table.schema(),
_query_request_projection(fts_req),
@@ -4577,13 +4577,13 @@ class AsyncTakeQuery(AsyncQueryBase):
def __init__(self, inner: LanceTakeQuery, table: Optional["AsyncTable"] = None):
super().__init__(inner, table)
def use_memwal(self, enable: bool) -> "AsyncTakeQuery":
"""Control MemWAL LSM read routing for this take query.
def use_rte(self, enable: bool) -> "AsyncTakeQuery":
"""Control RTE LSM read routing for this take query.
``False`` bypasses the MemWAL and reads the base table only — the escape
``False`` bypasses the RTE and reads the base table only — the escape
hatch, since take-by-row-id/offset is not supported on the LSM scanner.
"""
self._inner.use_memwal(enable)
self._inner.use_rte(enable)
return self
async def _plain_scan_to_pandas(
@@ -4644,14 +4644,14 @@ class BaseQueryBuilder(object):
self._inner.with_row_id()
return self
def use_memwal(self, enable: bool) -> Self:
def use_rte(self, enable: bool) -> Self:
"""
Control MemWAL LSM read routing for this query.
Control RTE LSM read routing for this query.
``False`` bypasses the MemWAL and reads the base table only, the escape
``False`` bypasses the RTE and reads the base table only, the escape
hatch for shapes the LSM scanner cannot honor (e.g. take-by-row-id).
"""
self._inner.use_memwal(enable)
self._inner.use_rte(enable)
return self
def with_row_address(self, with_row_address: bool = True) -> Self:
python/python/lancedb/remote/table.py
+24 -24
View File
@@ -31,7 +31,7 @@ from lancedb._lancedb import (
DropColumnsResult,
FtsToken,
IndexConfig,
MemWalWriteSpec,
RTEWriteSpec,
MergeResult,
UpdateResult,
)
@@ -1037,17 +1037,17 @@ class RemoteTable(Table):
"""Set the unenforced primary key for this table to a single column."""
return LOOP.run(self._table.set_unenforced_primary_key(columns))
def enable_memwal(self, spec: "MemWalWriteSpec") -> None:
"""Install an MemWalWriteSpec."""
return LOOP.run(self._table.enable_memwal(spec))
def enable_rte(self, spec: "RTEWriteSpec") -> None:
"""Install an RTEWriteSpec."""
return LOOP.run(self._table.enable_rte(spec))
def disable_memwal(self) -> None:
"""Remove the MemWalWriteSpec."""
return LOOP.run(self._table.disable_memwal())
def disable_rte(self) -> None:
"""Remove the RTEWriteSpec."""
return LOOP.run(self._table.disable_rte())
def get_memwal_spec(self) -> Optional["MemWalWriteSpec"]:
"""Read the installed MemWalWriteSpec, or ``None``."""
return LOOP.run(self._table.get_memwal_spec())
def get_rte_spec(self) -> Optional["RTEWriteSpec"]:
"""Read the installed RTEWriteSpec, or ``None``."""
return LOOP.run(self._table.get_rte_spec())
# WAL-PK-FUSION: delete both hooks.
def _hybrid_pk_fusion_learned(self) -> bool:
@@ -1057,31 +1057,31 @@ class RemoteTable(Table):
def _note_hybrid_pk_fusion(self) -> None:
self._table._note_hybrid_pk_fusion()
def checkpoint_memwal(self) -> None:
def checkpoint_rte(self) -> None:
"""Synchronous version of
[`AsyncTable.checkpoint_memwal`][lancedb.AsyncTable.checkpoint_memwal]."""
return LOOP.run(self._table.checkpoint_memwal())
[`AsyncTable.checkpoint_rte`][lancedb.AsyncTable.checkpoint_rte]."""
return LOOP.run(self._table.checkpoint_rte())
def flush_memwal(self) -> None:
def flush_rte(self) -> None:
"""Synchronous version of
[`AsyncTable.flush_memwal`][lancedb.AsyncTable.flush_memwal]."""
return LOOP.run(self._table.flush_memwal())
[`AsyncTable.flush_rte`][lancedb.AsyncTable.flush_rte]."""
return LOOP.run(self._table.flush_rte())
def compact_memwal(self) -> None:
def compact_rte(self) -> None:
"""Synchronous version of
[`AsyncTable.compact_memwal`][lancedb.AsyncTable.compact_memwal]."""
return LOOP.run(self._table.compact_memwal())
[`AsyncTable.compact_rte`][lancedb.AsyncTable.compact_rte]."""
return LOOP.run(self._table.compact_rte())
def get_memwal_stats(self, *, include_sstable_rows: bool = False) -> Optional[dict]:
def get_rte_stats(self, *, include_sstable_rows: bool = False) -> Optional[dict]:
"""Synchronous version of
[`AsyncTable.get_memwal_stats`][lancedb.AsyncTable.get_memwal_stats]."""
[`AsyncTable.get_rte_stats`][lancedb.AsyncTable.get_rte_stats]."""
return LOOP.run(
self._table.get_memwal_stats(include_sstable_rows=include_sstable_rows)
self._table.get_rte_stats(include_sstable_rows=include_sstable_rows)
)
def close_memwal_writers(self) -> None:
def close_rte_writers(self) -> None:
"""No-op for remote tables (no local shard writers)."""
return LOOP.run(self._table.close_memwal_writers())
return LOOP.run(self._table.close_rte_writers())
def drop_index(self, index_name: str):
return LOOP.run(self._table.drop_index(index_name))
python/python/lancedb/table.py
+72 -72
View File
@@ -223,7 +223,7 @@ if TYPE_CHECKING:
DeleteResult,
DropColumnsResult,
FtsToken,
MemWalWriteSpec,
RTEWriteSpec,
MergeResult,
UpdateResult,
)
@@ -4849,47 +4849,47 @@ class LanceTable(Table):
[`AsyncTable.set_unenforced_primary_key`][lancedb.AsyncTable.set_unenforced_primary_key]."""
return LOOP.run(self._table.set_unenforced_primary_key(columns))
def enable_memwal(self, spec: "MemWalWriteSpec") -> None:
"""Install an MemWalWriteSpec. See
[`AsyncTable.enable_memwal`][lancedb.AsyncTable.enable_memwal]."""
return LOOP.run(self._table.enable_memwal(spec))
def enable_rte(self, spec: "RTEWriteSpec") -> None:
"""Install an RTEWriteSpec. See
[`AsyncTable.enable_rte`][lancedb.AsyncTable.enable_rte]."""
return LOOP.run(self._table.enable_rte(spec))
def disable_memwal(self) -> None:
"""Remove the MemWalWriteSpec. See
[`AsyncTable.disable_memwal`][lancedb.AsyncTable.disable_memwal]."""
return LOOP.run(self._table.disable_memwal())
def disable_rte(self) -> None:
"""Remove the RTEWriteSpec. See
[`AsyncTable.disable_rte`][lancedb.AsyncTable.disable_rte]."""
return LOOP.run(self._table.disable_rte())
def get_memwal_spec(self) -> Optional["MemWalWriteSpec"]:
"""Read the installed MemWalWriteSpec, or ``None``. See
[`AsyncTable.get_memwal_spec`][lancedb.AsyncTable.get_memwal_spec]."""
return LOOP.run(self._table.get_memwal_spec())
def get_rte_spec(self) -> Optional["RTEWriteSpec"]:
"""Read the installed RTEWriteSpec, or ``None``. See
[`AsyncTable.get_rte_spec`][lancedb.AsyncTable.get_rte_spec]."""
return LOOP.run(self._table.get_rte_spec())
def checkpoint_memwal(self) -> None:
def checkpoint_rte(self) -> None:
"""Synchronous version of
[`AsyncTable.checkpoint_memwal`][lancedb.AsyncTable.checkpoint_memwal]."""
return LOOP.run(self._table.checkpoint_memwal())
[`AsyncTable.checkpoint_rte`][lancedb.AsyncTable.checkpoint_rte]."""
return LOOP.run(self._table.checkpoint_rte())
def flush_memwal(self) -> None:
def flush_rte(self) -> None:
"""Synchronous version of
[`AsyncTable.flush_memwal`][lancedb.AsyncTable.flush_memwal]."""
return LOOP.run(self._table.flush_memwal())
[`AsyncTable.flush_rte`][lancedb.AsyncTable.flush_rte]."""
return LOOP.run(self._table.flush_rte())
def compact_memwal(self) -> None:
def compact_rte(self) -> None:
"""Synchronous version of
[`AsyncTable.compact_memwal`][lancedb.AsyncTable.compact_memwal]."""
return LOOP.run(self._table.compact_memwal())
[`AsyncTable.compact_rte`][lancedb.AsyncTable.compact_rte]."""
return LOOP.run(self._table.compact_rte())
def get_memwal_stats(self, *, include_sstable_rows: bool = False) -> Optional[dict]:
def get_rte_stats(self, *, include_sstable_rows: bool = False) -> Optional[dict]:
"""Synchronous version of
[`AsyncTable.get_memwal_stats`][lancedb.AsyncTable.get_memwal_stats]."""
[`AsyncTable.get_rte_stats`][lancedb.AsyncTable.get_rte_stats]."""
return LOOP.run(
self._table.get_memwal_stats(include_sstable_rows=include_sstable_rows)
self._table.get_rte_stats(include_sstable_rows=include_sstable_rows)
)
def close_memwal_writers(self) -> None:
"""Close cached MemWAL shard writers. See
[`AsyncTable.close_memwal_writers`][lancedb.AsyncTable.close_memwal_writers]."""
return LOOP.run(self._table.close_memwal_writers())
def close_rte_writers(self) -> None:
"""Close cached RTE shard writers. See
[`AsyncTable.close_rte_writers`][lancedb.AsyncTable.close_rte_writers]."""
return LOOP.run(self._table.close_rte_writers())
def uses_v2_manifest_paths(self) -> bool:
"""
@@ -5547,24 +5547,24 @@ class AsyncTable:
columns = list(columns)
await self._inner.set_unenforced_primary_key(columns)
async def enable_memwal(self, spec: "MemWalWriteSpec") -> None:
"""Install an MemWalWriteSpec on this table.
async def enable_rte(self, spec: "RTEWriteSpec") -> None:
"""Install an RTEWriteSpec on this table.
The spec selects Lance's MemWAL LSM-style write path for future
`merge_insert` calls. ``MemWalWriteSpec`` chooses one of three sharding
`merge_insert` calls. ``RTEWriteSpec`` chooses one of three sharding
strategies:
- ``MemWalWriteSpec.bucket(column, num_buckets)`` — hash-bucket writes by
- ``RTEWriteSpec.bucket(column, num_buckets)`` — hash-bucket writes by
the single-column unenforced primary key.
- ``MemWalWriteSpec.identity(column)`` — shard by the raw value of a
- ``RTEWriteSpec.identity(column)`` — shard by the raw value of a
scalar column.
- ``MemWalWriteSpec.unsharded()`` — route every write to a single shard.
- ``RTEWriteSpec.unsharded()`` — route every write to a single shard.
All variants require the table to have an unenforced primary key set
via [`set_unenforced_primary_key`]; bucket sharding additionally
requires it to be the single column being bucketed.
By default the MemWAL maintains every index on the table, resolved
By default the RTE maintains every index on the table, resolved
here — a snapshot, so an index created afterwards needs the spec unset
and set again. This fails if one cannot be maintained; name the set
with ``with_maintained_indexes`` to install anyway. That pins an exact
@@ -5573,30 +5573,30 @@ class AsyncTable:
Parameters
----------
spec : MemWalWriteSpec
spec : RTEWriteSpec
The sharding spec to install.
Examples
--------
>>> from lancedb import MemWalWriteSpec
>>> from lancedb import RTEWriteSpec
>>> # table.set_unenforced_primary_key("id")
>>> # table.enable_memwal(MemWalWriteSpec.bucket("id", 16))
>>> # table.enable_rte(RTEWriteSpec.bucket("id", 16))
"""
await self._inner.enable_memwal(spec)
await self._inner.enable_rte(spec)
async def disable_memwal(self) -> None:
"""Remove the MemWalWriteSpec from this table.
async def disable_rte(self) -> None:
"""Remove the RTEWriteSpec from this table.
Reverts to the standard `merge_insert` write path. Errors if no spec
is currently set.
"""
await self._inner.disable_memwal()
await self._inner.disable_rte()
async def get_memwal_spec(self) -> Optional["MemWalWriteSpec"]:
"""Read the MemWalWriteSpec currently installed on this table.
async def get_rte_spec(self) -> Optional["RTEWriteSpec"]:
"""Read the RTEWriteSpec currently installed on this table.
Returns ``None`` when the MemWAL write path is not enabled at all — no
spec has been set, or one was removed with `disable_memwal`.
Returns ``None`` when the RTE write path is not enabled at all — no
spec has been set, or one was removed with `disable_rte`.
That is a different answer from a spec whose ``maintained_indexes``
is ``None``, which is an installed spec selecting indexes
automatically.
@@ -5606,7 +5606,7 @@ class AsyncTable:
or gains later, ``[]`` maintains none, and a non-empty list maintains
exactly those. All three round-trip.
"""
return await self._inner.get_memwal_spec()
return await self._inner.get_rte_spec()
# WAL-PK-FUSION: delete both hooks.
def _hybrid_pk_fusion_learned(self) -> bool:
@@ -5617,12 +5617,12 @@ class AsyncTable:
"""See [`Table._note_hybrid_pk_fusion`][lancedb.table.Table]."""
self._inner.note_hybrid_pk_fusion()
async def checkpoint_memwal(self) -> None:
"""Converge this table's MemWAL write path into its base table.
async def checkpoint_rte(self) -> None:
"""Converge this table's RTE write path into its base table.
One flush, freezing every memtable into an SSTable, then compaction triggers
until every SSTable that existed at that moment has reached base.
The loop runs client-side, reading progress from ``get_memwal_stats``.
The loop runs client-side, reading progress from ``get_rte_stats``.
Best-effort: SSTables created *while* it runs are deliberately not
waited on, which is what lets it terminate on a table taking writes.
@@ -5637,56 +5637,56 @@ class AsyncTable:
``asyncio.wait_for`` for a wall-clock bound; abandoning it partway
costs nothing.
"""
await self._inner.checkpoint_memwal()
await self._inner.checkpoint_rte()
async def flush_memwal(self) -> None:
async def flush_rte(self) -> None:
"""Freeze every table shard's active memtable into an SSTable.
Does not touch the base table — compacting SSTables into base is
`compact_memwal`. On a node that has not claimed this table, this claims
`compact_rte`. On a node that has not claimed this table, this claims
it and replays its WAL log first.
"""
await self._inner.flush_memwal()
await self._inner.flush_rte()
async def compact_memwal(self) -> None:
async def compact_rte(self) -> None:
"""Trigger a background SSTable compaction pass per table shard.
Returns once the passes are dispatched, not once they finish: watch
``get_memwal_stats`` for progress, or use ``checkpoint_memwal`` to loop
``get_rte_stats`` for progress, or use ``checkpoint_rte`` to loop
until the current SSTables have reached base.
"""
await self._inner.compact_memwal()
await self._inner.compact_rte()
async def get_memwal_stats(
async def get_rte_stats(
self, *, include_sstable_rows: bool = False
) -> Optional[dict]:
"""Read live per-table-shard MemWAL state.
"""Read live per-table-shard RTE state.
Answers "how far behind is my fresh tier", "which table shard is hot", and
"why is my fresh-tier vector search brute-force". Mutates no table
state, though on a node that has not claimed this table it claims it,
exactly as a read would.
Returns ``None`` only when the MemWAL write path is not enabled.
Returns ``None`` only when the RTE write path is not enabled.
Parameters
----------
include_sstable_rows
Report a row count per SSTable. Off by default: each count
opens an uncached Lance dataset, and ``checkpoint_memwal`` polls this
opens an uncached Lance dataset, and ``checkpoint_rte`` polls this
needing only generation numbers.
"""
return await self._inner.get_memwal_stats(include_sstable_rows)
return await self._inner.get_rte_stats(include_sstable_rows)
async def close_memwal_writers(self) -> None:
"""Drain and close any cached MemWAL shard writers for this table.
async def close_rte_writers(self) -> None:
"""Drain and close any cached RTE shard writers for this table.
When a MemWAL write spec is installed, `merge_insert` opens MemWAL shard
When an RTE write spec is installed, `merge_insert` opens RTE shard
writers and caches them for reuse across calls. This closes them,
flushing pending data; writers reopen lazily on the next
`merge_insert`. It is a no-op when no writers are cached.
"""
await self._inner.close_memwal_writers()
await self._inner.close_rte_writers()
@property
def name(self) -> str:
@@ -6575,8 +6575,8 @@ class AsyncTable:
take_query = self.take_offsets(query.take_offsets)
if query.columns:
take_query = take_query.select(query.columns)
if query.use_memwal is not None:
take_query = take_query.use_memwal(query.use_memwal)
if query.use_rte is not None:
take_query = take_query.use_rte(query.use_rte)
if query.with_row_id:
take_query = take_query.with_row_id()
return take_query
@@ -6592,8 +6592,8 @@ class AsyncTable:
async_query = async_query.where(query.filter)
if query.fast_search:
async_query = async_query.fast_search()
if query.use_memwal is not None:
async_query = async_query.use_memwal(query.use_memwal)
if query.use_rte is not None:
async_query = async_query.use_rte(query.use_rte)
if query.with_row_id:
async_query = async_query.with_row_id()
if query.order_by:
@@ -6719,7 +6719,7 @@ class AsyncTable:
when_not_matched_by_source_condition_expr=merge._when_not_matched_by_source_condition_expr,
timeout=merge._timeout,
use_index=merge._use_index,
use_memwal=merge._use_memwal,
use_rte=merge._use_rte,
validate_single_shard=merge._validate_single_shard,
),
)
python/python/tests/test_merge_insert_lsm.py
+50 -56
View File
@@ -1,14 +1,14 @@
# SPDX-License-Identifier: Apache-2.0
# SPDX-FileCopyrightText: Copyright The LanceDB Authors
"""Tests for the MemWAL LSM ``merge_insert`` dispatch."""
"""Tests for the RTE LSM ``merge_insert`` dispatch."""
from datetime import timedelta
import lancedb
import pyarrow as pa
import pytest
from lancedb._lancedb import MemWalWriteSpec
from lancedb._lancedb import RTEWriteSpec
from lancedb.index import FTS, IvfPq
SCHEMA = pa.schema(
@@ -49,12 +49,12 @@ def _region_reader(rows):
def _bucket_table(tmp_path):
"""A table with ``id`` as the primary key and a single-bucket MemWAL spec."""
"""A table with ``id`` as the primary key and a single-bucket RTE spec."""
db = lancedb.connect(tmp_path, read_consistency_interval=timedelta(seconds=0))
table = db.create_table("t", _reader([1, 2, 3]))
table.set_unenforced_primary_key("id")
# num_buckets = 1: every row routes to the single bucket.
table.enable_memwal(MemWalWriteSpec.bucket("id", 1))
table.enable_rte(RTEWriteSpec.bucket("id", 1))
return table
@@ -67,7 +67,7 @@ def test_lsm_merge_insert_bucket(tmp_path):
.when_not_matched_insert_all()
.execute(_reader([3, 4, 5]))
)
# LSM path: rows go to the MemWAL, so only num_rows is populated.
# LSM path: rows go to the RTE, so only num_rows is populated.
assert result.num_rows == 3
assert result.version == 0
assert result.num_inserted_rows == 0
@@ -78,7 +78,7 @@ def test_lsm_merge_insert_unsharded(tmp_path):
db = lancedb.connect(tmp_path, read_consistency_interval=timedelta(seconds=0))
table = db.create_table("t", _reader([1, 2, 3]))
table.set_unenforced_primary_key("id")
table.enable_memwal(MemWalWriteSpec.unsharded())
table.enable_rte(RTEWriteSpec.unsharded())
result = (
table.merge_insert("id")
.when_matched_update_all()
@@ -92,7 +92,7 @@ def test_lsm_merge_insert_identity(tmp_path):
db = lancedb.connect(tmp_path, read_consistency_interval=timedelta(seconds=0))
table = db.create_table("t", _region_reader([(1, "us"), (2, "us")]))
table.set_unenforced_primary_key("id")
table.enable_memwal(MemWalWriteSpec.identity("region"))
table.enable_rte(RTEWriteSpec.identity("region"))
# All rows share one identity value, so they route to one shard.
result = (
table.merge_insert([])
@@ -103,31 +103,31 @@ def test_lsm_merge_insert_identity(tmp_path):
assert result.num_rows == 2
def test_lsm_merge_insert_use_memwal_false(tmp_path):
def test_lsm_merge_insert_use_rte_false(tmp_path):
table = _bucket_table(tmp_path) # rows id = 1, 2, 3
# use_memwal(False) opts out: the standard path runs and commits even with a spec.
# use_rte(False) opts out: the standard path runs and commits even with a spec.
result = (
table.merge_insert("id")
.when_not_matched_insert_all()
.use_memwal(False)
.use_rte(False)
.execute(_reader([3, 4, 5]))
)
assert result.num_inserted_rows == 2
assert table.count_rows() == 5
def test_lsm_merge_insert_use_memwal_true_without_spec_errors(tmp_path):
# A table with a primary key but no MemWAL write spec installed.
def test_lsm_merge_insert_use_rte_true_without_spec_errors(tmp_path):
# A table with a primary key but no RTE write spec installed.
db = lancedb.connect(tmp_path, read_consistency_interval=timedelta(seconds=0))
table = db.create_table("t", _reader([1, 2, 3]))
table.set_unenforced_primary_key("id")
# use_memwal(True) demands MemWAL routing; without a spec it errors.
with pytest.raises(Exception, match="use_memwal"):
# use_rte(True) demands RTE routing; without a spec it errors.
with pytest.raises(Exception, match="use_rte"):
(
table.merge_insert("id")
.when_matched_update_all()
.when_not_matched_insert_all()
.use_memwal(True)
.use_rte(True)
.execute(_reader([3, 4, 5]))
)
@@ -145,7 +145,7 @@ def test_lsm_merge_insert_validate_single_shard_off(tmp_path):
def test_lsm_merge_insert_no_spec_uses_standard_path(tmp_path):
# A table with a primary key but no MemWAL write spec installed.
# A table with a primary key but no RTE write spec installed.
db = lancedb.connect(tmp_path, read_consistency_interval=timedelta(seconds=0))
table = db.create_table("t", _reader([1, 2, 3]))
table.set_unenforced_primary_key("id")
@@ -186,7 +186,7 @@ def test_lsm_close_writers(tmp_path):
.when_not_matched_insert_all()
.execute(_reader([7, 8]))
)
table.close_memwal_writers()
table.close_rte_writers()
# The writer reopens lazily on the next merge_insert.
result = (
table.merge_insert([])
@@ -204,14 +204,14 @@ async def test_async_lsm_merge_insert(tmp_path):
)
table = await db.create_table("t", _reader([1, 2, 3]))
await table.set_unenforced_primary_key("id")
await table.enable_memwal(MemWalWriteSpec.bucket("id", 1))
await table.enable_rte(RTEWriteSpec.bucket("id", 1))
builder = (
table.merge_insert([]).when_matched_update_all().when_not_matched_insert_all()
)
result = await builder.execute(_reader([3, 4, 5]))
assert result.num_rows == 3
await table.close_memwal_writers()
await table.close_rte_writers()
def _lsm_upsert(table, ids):
@@ -228,7 +228,7 @@ def test_lsm_read_sees_active_memtable(tmp_path):
db = lancedb.connect(tmp_path, read_consistency_interval=timedelta(seconds=0))
table = db.create_table("t", _reader([1, 2, 3])) # base ids 1,2,3
table.set_unenforced_primary_key("id")
table.enable_memwal(MemWalWriteSpec.unsharded())
table.enable_rte(RTEWriteSpec.unsharded())
_lsm_upsert(table, [4, 5]) # active memtable only, not committed to base
@@ -236,8 +236,8 @@ def test_lsm_read_sees_active_memtable(tmp_path):
lsm = table.search().to_arrow()
assert sorted(lsm["id"].to_pylist()) == [1, 2, 3, 4, 5]
# use_memwal(False) bypasses the MemWAL and reads the base table only.
base_only = table.search().use_memwal(False).to_arrow()
# use_rte(False) bypasses the RTE and reads the base table only.
base_only = table.search().use_rte(False).to_arrow()
assert sorted(base_only["id"].to_pylist()) == [1, 2, 3]
@@ -245,7 +245,7 @@ def test_lsm_read_dedup_newest_wins(tmp_path):
db = lancedb.connect(tmp_path, read_consistency_interval=timedelta(seconds=0))
table = db.create_table("t", _reader([1, 2, 3])) # id 2 -> value 1
table.set_unenforced_primary_key("id")
table.enable_memwal(MemWalWriteSpec.unsharded())
table.enable_rte(RTEWriteSpec.unsharded())
_lsm_upsert(table, [2, 3, 4]) # ids 2,3,4 -> values 0,1,2
@@ -258,31 +258,31 @@ def test_lsm_read_dedup_newest_wins(tmp_path):
def test_lsm_read_without_spec_reads_base(tmp_path):
db = lancedb.connect(tmp_path, read_consistency_interval=timedelta(seconds=0))
table = db.create_table("t", _reader([1, 2, 3]))
table.set_unenforced_primary_key("id") # no MemWAL write spec
table.set_unenforced_primary_key("id") # no RTE write spec
# No spec: default read and use_memwal(False) both read the base table, no error.
# No spec: default read and use_rte(False) both read the base table, no error.
assert sorted(table.search().to_arrow()["id"].to_pylist()) == [1, 2, 3]
assert sorted(table.search().use_memwal(False).to_arrow()["id"].to_pylist()) == [
assert sorted(table.search().use_rte(False).to_arrow()["id"].to_pylist()) == [
1,
2,
3,
]
def test_lsm_read_unsupported_shape_errors_without_use_memwal_false(tmp_path):
def test_lsm_read_unsupported_shape_errors_without_use_rte_false(tmp_path):
db = lancedb.connect(tmp_path, read_consistency_interval=timedelta(seconds=0))
table = db.create_table("t", _reader([1, 2, 3]))
table.set_unenforced_primary_key("id")
table.enable_memwal(MemWalWriteSpec.unsharded())
table.enable_rte(RTEWriteSpec.unsharded())
_lsm_upsert(table, [4])
# with_row_id is unsupported by the LSM scanner; on a MemWAL table the default
# with_row_id is unsupported by the LSM scanner; on an RTE table the default
# (auto-routed) read hard-errors instead of silently reading a stale base.
with pytest.raises(Exception):
table.search().with_row_id(True).to_arrow()
# use_memwal(False) is the escape hatch: it reads the base table only.
base = table.search().with_row_id(True).use_memwal(False).to_arrow()
# use_rte(False) is the escape hatch: it reads the base table only.
base = table.search().with_row_id(True).use_rte(False).to_arrow()
assert sorted(base["id"].to_pylist()) == [1, 2, 3]
@@ -293,7 +293,7 @@ async def test_async_lsm_read(tmp_path):
)
table = await db.create_table("t", _reader([1, 2, 3]))
await table.set_unenforced_primary_key("id")
await table.enable_memwal(MemWalWriteSpec.unsharded())
await table.enable_rte(RTEWriteSpec.unsharded())
builder = (
table.merge_insert([]).when_matched_update_all().when_not_matched_insert_all()
@@ -331,7 +331,7 @@ def _vector_reader(rows):
def _vector_table(tmp_path):
"""Base table whose vector column is indexed so its rows are visible to the LSM
vector scanner (the base arm uses ``fast_search`` — indexed data only), plus an
unsharded MemWAL spec that maintains that index for the memtable.
unsharded RTE spec that maintains that index for the memtable.
Rows 1,2 are category ``a``, row 3 is ``b``, and 4..60 are filler ``c`` that
give the tiny IVF index enough data to train.
@@ -353,9 +353,7 @@ def _vector_table(tmp_path):
"vector", config=IvfPq(num_partitions=1, num_sub_vectors=2, num_bits=4)
)
index_name = table.list_indices()[0].name
table.enable_memwal(
MemWalWriteSpec.unsharded().with_maintained_indexes([index_name])
)
table.enable_rte(RTEWriteSpec.unsharded().with_maintained_indexes([index_name]))
return table
@@ -379,9 +377,9 @@ def test_lsm_read_vector_sees_memtable(tmp_path):
assert {1, 2, 3} <= ids # indexed base rows
assert 1000 in ids # in-flight memtable row
# use_memwal(False) bypasses the MemWAL, so the in-flight row is not visible.
# use_rte(False) bypasses the RTE, so the in-flight row is not visible.
base_ids = set(
table.search(query).use_memwal(False).limit(100).to_arrow()["id"].to_pylist()
table.search(query).use_rte(False).limit(100).to_arrow()["id"].to_pylist()
)
assert {1, 2, 3} <= base_ids
assert 1000 not in base_ids
@@ -453,9 +451,7 @@ def test_lsm_read_fts_sees_memtable(tmp_path):
# Native FTS index (tantivy is not compatible with the LSM memtable index).
table.create_index("text", config=FTS())
index_name = table.list_indices()[0].name
table.enable_memwal(
MemWalWriteSpec.unsharded().with_maintained_indexes([index_name])
)
table.enable_rte(RTEWriteSpec.unsharded().with_maintained_indexes([index_name]))
# in-flight doc 4 lands in the memtable's maintained FTS index.
(
@@ -498,8 +494,8 @@ def test_lsm_read_vector_unsupported_knobs_error(tmp_path):
with pytest.raises(Exception, match="use_index"):
table.search(query).bypass_vector_index().to_arrow()
# use_memwal(False) is the escape hatch: the base-only standard path honors them.
base = table.search(query).distance_range(0.0, 100.0).use_memwal(False).to_arrow()
# use_rte(False) is the escape hatch: the base-only standard path honors them.
base = table.search(query).distance_range(0.0, 100.0).use_rte(False).to_arrow()
assert 1000 not in set(base["id"].to_pylist())
@@ -544,7 +540,7 @@ def test_lsm_read_fts_unmaintained_index_errors(tmp_path):
table.create_index("text", config=FTS())
# No maintained indexes: the active memtable FTS arm cannot serve un-compacted
# docs, so the search would silently omit them — reject instead.
table.enable_memwal(MemWalWriteSpec.unsharded().with_maintained_indexes([]))
table.enable_rte(RTEWriteSpec.unsharded().with_maintained_indexes([]))
with pytest.raises(Exception, match="maintained"):
table.search("fox", query_type="fts", fts_columns="text").to_arrow()
@@ -553,7 +549,7 @@ def test_lsm_read_time_travel_errors(tmp_path):
db = lancedb.connect(tmp_path, read_consistency_interval=timedelta(seconds=0))
table = db.create_table("t", _reader([1, 2, 3]))
table.set_unenforced_primary_key("id")
table.enable_memwal(MemWalWriteSpec.unsharded())
table.enable_rte(RTEWriteSpec.unsharded())
pinned = table.version
table.add(_reader([4, 5])) # standard add commits a newer version
table.checkout(pinned) # detached head at the historical version
@@ -562,8 +558,8 @@ def test_lsm_read_time_travel_errors(tmp_path):
# historical version is rejected.
with pytest.raises(Exception, match="time-travel"):
table.search().to_arrow()
# use_memwal(False) reads the base table at the pinned version.
base = table.search().use_memwal(False).to_arrow()
# use_rte(False) reads the base table at the pinned version.
base = table.search().use_rte(False).to_arrow()
assert sorted(base["id"].to_pylist()) == [1, 2, 3]
@@ -571,14 +567,14 @@ def test_lsm_read_take_row_ids_errors(tmp_path):
db = lancedb.connect(tmp_path, read_consistency_interval=timedelta(seconds=0))
table = db.create_table("t", _reader([1, 2, 3]))
table.set_unenforced_primary_key("id")
table.enable_memwal(MemWalWriteSpec.unsharded())
table.enable_rte(RTEWriteSpec.unsharded())
_lsm_upsert(table, [4])
# take-by-row-id auto-routes through the LSM scanner, which has no stable _rowid,
# so it hard-errors instead of failing with an opaque column-not-found error.
with pytest.raises(Exception, match="row id"):
table.take_row_ids([0, 1]).to_arrow()
# use_memwal(False) is the escape hatch: it reads the base table.
base = table.take_row_ids([0, 1]).use_memwal(False).to_arrow()
# use_rte(False) is the escape hatch: it reads the base table.
base = table.take_row_ids([0, 1]).use_rte(False).to_arrow()
assert base.num_rows == 2
@@ -588,9 +584,7 @@ def test_lsm_read_fts_postfilter_errors(tmp_path):
table.set_unenforced_primary_key("id")
table.create_index("text", config=FTS())
index_name = table.list_indices()[0].name
table.enable_memwal(
MemWalWriteSpec.unsharded().with_maintained_indexes([index_name])
)
table.enable_rte(RTEWriteSpec.unsharded().with_maintained_indexes([index_name]))
# The LSM scanner always prefilters; postfilter on FTS changes result semantics,
# so it hard-errors (previously only the vector arm rejected it).
with pytest.raises(Exception, match="postfilter"):
@@ -608,7 +602,7 @@ def test_lsm_read_fts_multiple_same_type_indexes_errors(tmp_path):
table.set_unenforced_primary_key("id")
table.create_index("text", config=FTS(), name="fts_a")
table.create_index("text", config=FTS(), name="fts_b", replace=False)
table.enable_memwal(MemWalWriteSpec.unsharded().with_maintained_indexes(["fts_a"]))
table.enable_rte(RTEWriteSpec.unsharded().with_maintained_indexes(["fts_a"]))
# Two FTS indexes on the column: the base planner's chosen index is ambiguous, so
# the scanner can't pick a catch-up watermark and rejects rather than risk
# dropping rows the actually-used index has not caught up to.
@@ -629,7 +623,7 @@ def test_lsm_read_vector_unmaintained_index_errors(tmp_path):
)
# Spec with NO maintained indexes: the base vector index's catch-up is untracked,
# so the scanner rejects rather than risk dropping compacted-but-unindexed rows.
table.enable_memwal(MemWalWriteSpec.unsharded().with_maintained_indexes([]))
table.enable_rte(RTEWriteSpec.unsharded().with_maintained_indexes([]))
with pytest.raises(Exception, match="maintained"):
table.search([1.0] * VECTOR_DIM).to_arrow()
@@ -642,7 +636,7 @@ def test_lsm_read_fts_optimized_index_not_rejected(tmp_path):
table.add(_fts_reader([(i, "lazy fox") for i in range(6, 11)]))
table.optimize() # may split the FTS index into multiple physical segments
name = table.list_indices()[0].name
table.enable_memwal(MemWalWriteSpec.unsharded().with_maintained_indexes([name]))
table.enable_rte(RTEWriteSpec.unsharded().with_maintained_indexes([name]))
# Multiple physical segments of one logical index must not be miscounted as
# multiple indexes and rejected.
ids = set(
python/python/tests/test_remote_db.py
+20 -20
View File
@@ -1844,10 +1844,10 @@ def lsm_test_table(lsm_handler):
"""A remote table whose LSM routes are served by ``lsm_handler``.
``lsm_handler(request, route)`` is called for ``/v1/table/test/<route>/``
where route is one of flush_memwal, compact_memwal, get_memwal_stats, and is
where route is one of flush_rte, compact_rte, get_rte_stats, and is
responsible for writing the response.
"""
routes = ("flush_memwal", "compact_memwal", "get_memwal_stats")
routes = ("flush_rte", "compact_rte", "get_rte_stats")
def handler(request):
match = re.fullmatch(r"/v1/table/test/(\w+)/", request.path)
@@ -1879,7 +1879,7 @@ def send_json(request, payload, status=200):
request.wfile.write(json.dumps(payload).encode())
def test_get_memwal_stats_sync():
def test_get_rte_stats_sync():
"""The sync wrapper round-trips the server payload into a dict."""
table_shard = {
"shard_id": "b0",
@@ -1904,29 +1904,29 @@ def test_get_memwal_stats_sync():
seen_bodies = []
def lsm_handler(request, route):
assert route == "get_memwal_stats"
assert route == "get_rte_stats"
seen_bodies.append(read_json_body(request))
send_json(request, {"memwal_stats": {"table_shards": [table_shard]}})
send_json(request, {"rte_stats": {"table_shards": [table_shard]}})
with lsm_test_table(lsm_handler) as table:
assert table.get_memwal_stats() == {"table_shards": [table_shard]}
assert table.get_rte_stats() == {"table_shards": [table_shard]}
# Off by default, and forwarded when asked for.
assert seen_bodies == [{"include_sstable_rows": False}]
table.get_memwal_stats(include_sstable_rows=True)
table.get_rte_stats(include_sstable_rows=True)
assert seen_bodies[-1] == {"include_sstable_rows": True}
def test_get_memwal_stats_sync_returns_none_when_memwal_disabled():
"""A null envelope means the MemWAL write path is not enabled, not an error."""
def test_get_rte_stats_sync_returns_none_when_rte_disabled():
"""A null envelope means the RTE write path is not enabled, not an error."""
def lsm_handler(request, route):
send_json(request, {"memwal_stats": None})
send_json(request, {"rte_stats": None})
with lsm_test_table(lsm_handler) as table:
assert table.get_memwal_stats() is None
assert table.get_rte_stats() is None
def test_flush_and_compact_memwal_sync():
def test_flush_and_compact_rte_sync():
"""Both are one-shot POSTs answered 202 with no body."""
called = []
@@ -1936,12 +1936,12 @@ def test_flush_and_compact_memwal_sync():
request.end_headers()
with lsm_test_table(lsm_handler) as table:
assert table.flush_memwal() is None
assert table.compact_memwal() is None
assert called == ["flush_memwal", "compact_memwal"]
assert table.flush_rte() is None
assert table.compact_rte() is None
assert called == ["flush_rte", "compact_rte"]
def test_checkpoint_memwal_sync():
def test_checkpoint_rte_sync():
"""Freeze, read the watermark, and return once no SSTables remain.
The convergence loop itself is covered in Rust; this pins the sync
@@ -1951,17 +1951,17 @@ def test_checkpoint_memwal_sync():
def lsm_handler(request, route):
called.append(route)
if route == "get_memwal_stats":
if route == "get_rte_stats":
# An empty SSTable tier yields no target watermark, so the loop is done
# after the freeze without ever polling compaction.
send_json(request, {"memwal_stats": {"table_shards": []}})
send_json(request, {"rte_stats": {"table_shards": []}})
else:
request.send_response(202)
request.end_headers()
with lsm_test_table(lsm_handler) as table:
assert table.checkpoint_memwal() is None
assert called == ["flush_memwal", "get_memwal_stats"]
assert table.checkpoint_rte() is None
assert called == ["flush_rte", "get_rte_stats"]
@contextlib.contextmanager
python/python/tests/test_memwal_spec.py → python/python/tests/test_rte_spec.py
+60 -60
View File
@@ -1,8 +1,8 @@
# SPDX-License-Identifier: Apache-2.0
# SPDX-FileCopyrightText: Copyright The LanceDB Authors
"""Tests for installing and clearing an MemWalWriteSpec via
`Table.enable_memwal` / `Table.disable_memwal`.
"""Tests for installing and clearing an RTEWriteSpec via
`Table.enable_rte` / `Table.disable_rte`.
"""
from datetime import timedelta
@@ -10,7 +10,7 @@ from datetime import timedelta
import lancedb
import pyarrow as pa
import pytest
from lancedb._lancedb import MemWalWriteSpec
from lancedb._lancedb import RTEWriteSpec
from lancedb.index import BTree
SCHEMA = pa.schema(
@@ -21,9 +21,9 @@ SCHEMA = pa.schema(
)
def test_memwal_spec_module_metadata():
assert lancedb.MemWalWriteSpec is MemWalWriteSpec
assert MemWalWriteSpec.__module__ == "lancedb._lancedb"
def test_rte_spec_module_metadata():
assert lancedb.RTEWriteSpec is RTEWriteSpec
assert RTEWriteSpec.__module__ == "lancedb._lancedb"
def _batch(ids, vs):
@@ -43,49 +43,49 @@ def _make_table(tmp_path):
return db, table
def test_enable_memwal_validates(tmp_path):
def test_enable_rte_validates(tmp_path):
_db, table = _make_table(tmp_path)
table.set_unenforced_primary_key("id")
# Out-of-range num_buckets.
with pytest.raises(Exception, match="num_buckets"):
table.enable_memwal(MemWalWriteSpec.bucket("id", 0))
table.enable_rte(RTEWriteSpec.bucket("id", 0))
with pytest.raises(Exception, match="num_buckets"):
table.enable_memwal(MemWalWriteSpec.bucket("id", 1025))
table.enable_rte(RTEWriteSpec.bucket("id", 1025))
# Happy path, then a set over the installed spec is refused.
table.enable_memwal(MemWalWriteSpec.bucket("id", 4))
table.enable_rte(RTEWriteSpec.bucket("id", 4))
with pytest.raises(Exception, match="already set"):
table.enable_memwal(MemWalWriteSpec.bucket("id", 8))
table.enable_rte(RTEWriteSpec.bucket("id", 8))
def test_disable_memwal(tmp_path):
def test_disable_rte(tmp_path):
_db, table = _make_table(tmp_path)
table.set_unenforced_primary_key("id")
# unset errors when no spec is set.
with pytest.raises(Exception, match="no MemWAL write spec"):
table.disable_memwal()
with pytest.raises(Exception, match="no RTE write spec"):
table.disable_rte()
# Install a spec, then remove it; afterwards a fresh spec can be set.
table.enable_memwal(MemWalWriteSpec.bucket("id", 4))
table.disable_memwal()
table.enable_rte(RTEWriteSpec.bucket("id", 4))
table.disable_rte()
# A second unset errors — there is no spec left to remove.
with pytest.raises(Exception, match="no MemWAL write spec"):
table.disable_memwal()
table.enable_memwal(MemWalWriteSpec.bucket("id", 8))
with pytest.raises(Exception, match="no RTE write spec"):
table.disable_rte()
table.enable_rte(RTEWriteSpec.bucket("id", 8))
def test_enable_memwal_requires_primary_key(tmp_path):
def test_enable_rte_requires_primary_key(tmp_path):
_db, table = _make_table(tmp_path)
for spec in (
MemWalWriteSpec.unsharded(),
MemWalWriteSpec.bucket("id", 4),
MemWalWriteSpec.identity("id"),
RTEWriteSpec.unsharded(),
RTEWriteSpec.bucket("id", 4),
RTEWriteSpec.identity("id"),
):
with pytest.raises(Exception, match="primary key"):
table.enable_memwal(spec)
assert table.get_memwal_spec() is None
table.enable_rte(spec)
assert table.get_rte_spec() is None
def test_set_unsharded_spec(tmp_path):
@@ -93,12 +93,12 @@ def test_set_unsharded_spec(tmp_path):
# Lance MemWAL still requires a primary key on the dataset; Unsharded
# just skips per-row hashing.
table.set_unenforced_primary_key("id")
table.enable_memwal(MemWalWriteSpec.unsharded())
table.disable_memwal()
table.enable_rte(RTEWriteSpec.unsharded())
table.disable_rte()
def test_memwal_spec_repr():
s = MemWalWriteSpec.bucket("id", 4)
def test_rte_spec_repr():
s = RTEWriteSpec.bucket("id", 4)
assert s.spec_type == "bucket"
assert s.column == "id"
assert s.num_buckets == 4
@@ -109,20 +109,20 @@ def test_memwal_spec_repr():
assert "id" in repr(s)
assert "4" in repr(s)
u = MemWalWriteSpec.unsharded()
u = RTEWriteSpec.unsharded()
assert u.spec_type == "unsharded"
assert u.column is None
assert u.num_buckets is None
assert "unsharded" in repr(u)
def test_memwal_spec_with_maintained_indexes():
s = MemWalWriteSpec.bucket("id", 4).with_maintained_indexes(["idx_a", "idx_b"])
def test_rte_spec_with_maintained_indexes():
s = RTEWriteSpec.bucket("id", 4).with_maintained_indexes(["idx_a", "idx_b"])
assert s.maintained_indexes == ["idx_a", "idx_b"]
@pytest.mark.asyncio
async def test_async_set_disable_memwal(tmp_path):
async def test_async_set_disable_rte(tmp_path):
db = await lancedb.connect_async(
tmp_path, read_consistency_interval=timedelta(seconds=0)
)
@@ -132,11 +132,11 @@ async def test_async_set_disable_memwal(tmp_path):
)
await table.set_unenforced_primary_key("id")
await table.enable_memwal(MemWalWriteSpec.bucket("id", 4))
await table.disable_memwal()
await table.enable_rte(RTEWriteSpec.bucket("id", 4))
await table.disable_rte()
# A second unset errors.
with pytest.raises(Exception, match="no MemWAL write spec"):
await table.disable_memwal()
with pytest.raises(Exception, match="no RTE write spec"):
await table.disable_rte()
def test_set_identity_spec(tmp_path):
@@ -144,12 +144,12 @@ def test_set_identity_spec(tmp_path):
# Identity sharding still requires an unenforced primary key on the
# table; it shards by the raw value of the given column.
table.set_unenforced_primary_key("id")
table.enable_memwal(MemWalWriteSpec.identity("v"))
table.disable_memwal()
table.enable_rte(RTEWriteSpec.identity("v"))
table.disable_rte()
def test_memwal_spec_identity_and_writer_config_defaults():
s = MemWalWriteSpec.identity("v")
def test_rte_spec_identity_and_writer_config_defaults():
s = RTEWriteSpec.identity("v")
assert s.spec_type == "identity"
assert s.column == "v"
assert s.num_buckets is None
@@ -160,12 +160,12 @@ def test_memwal_spec_identity_and_writer_config_defaults():
assert "durable_write" in repr(s)
def test_get_memwal_spec(tmp_path):
def test_get_rte_spec(tmp_path):
_db, table = _make_table(tmp_path)
table.set_unenforced_primary_key("id")
# None when nothing is installed.
assert table.get_memwal_spec() is None
assert table.get_rte_spec() is None
# A real scalar index is needed to name it as a maintained index.
table.create_index("id", config=BTree())
@@ -173,12 +173,12 @@ def test_get_memwal_spec(tmp_path):
# Bucket spec round-trips, including maintained indexes and writer config
# defaults.
table.enable_memwal(
MemWalWriteSpec.bucket("id", 4)
table.enable_rte(
RTEWriteSpec.bucket("id", 4)
.with_maintained_indexes([idx_name])
.with_writer_config_defaults({"durable_write": "false"})
)
spec = table.get_memwal_spec()
spec = table.get_rte_spec()
assert spec is not None
assert spec.spec_type == "bucket"
assert spec.column == "id"
@@ -187,30 +187,30 @@ def test_get_memwal_spec(tmp_path):
assert spec.writer_config_defaults == {"durable_write": "false"}
# After unset, None again.
table.disable_memwal()
assert table.get_memwal_spec() is None
table.disable_rte()
assert table.get_rte_spec() is None
# Identity round-trips (column recovered from the schema). An unnamed
# maintained set reads back as None: what is stored is the intent to
# maintain every index, not the set it resolves to now.
table.enable_memwal(MemWalWriteSpec.identity("id"))
spec = table.get_memwal_spec()
table.enable_rte(RTEWriteSpec.identity("id"))
spec = table.get_rte_spec()
assert spec.spec_type == "identity"
assert spec.column == "id"
assert spec.maintained_indexes is None
table.disable_memwal()
table.disable_rte()
# Unsharded round-trips (no routing column). Opting out is distinct from
# the unnamed default.
table.enable_memwal(MemWalWriteSpec.unsharded().with_maintained_indexes([]))
spec = table.get_memwal_spec()
table.enable_rte(RTEWriteSpec.unsharded().with_maintained_indexes([]))
spec = table.get_rte_spec()
assert spec.spec_type == "unsharded"
assert spec.column is None
assert spec.maintained_indexes == []
@pytest.mark.asyncio
async def test_async_get_memwal_spec(tmp_path):
async def test_async_get_rte_spec(tmp_path):
db = await lancedb.connect_async(
tmp_path, read_consistency_interval=timedelta(seconds=0)
)
@@ -219,21 +219,21 @@ async def test_async_get_memwal_spec(tmp_path):
pa.RecordBatchReader.from_batches(SCHEMA, [_batch(["seed"], [0])]),
)
assert await table.get_memwal_spec() is None
assert await table.get_rte_spec() is None
await table.set_unenforced_primary_key("id")
# A real scalar index is needed to name it as a maintained index.
await table.create_index("id", config=BTree())
idx_name = (await table.list_indices())[0].name
await table.enable_memwal(
MemWalWriteSpec.bucket("id", 8).with_maintained_indexes([idx_name])
await table.enable_rte(
RTEWriteSpec.bucket("id", 8).with_maintained_indexes([idx_name])
)
spec = await table.get_memwal_spec()
spec = await table.get_rte_spec()
assert spec is not None
assert spec.spec_type == "bucket"
assert spec.column == "id"
assert spec.num_buckets == 8
assert spec.maintained_indexes == [idx_name]
await table.disable_memwal()
assert await table.get_memwal_spec() is None
await table.disable_rte()
assert await table.get_rte_spec() is None
python/python/tests/test_wal_hybrid.py
+7 -7
View File
@@ -14,7 +14,7 @@ from lancedb.rerankers import RRFReranker
def _wal_table(schema: pa.Schema) -> mock.Mock:
"""A MemWAL-backed table a previous query has already been refused
"""An RTE-backed table a previous query has already been refused
`_rowid` on, so the fusion goes straight to the primary key."""
table = mock.Mock()
table._hybrid_pk_fusion_learned.return_value = True
@@ -51,7 +51,7 @@ def test_wal_hybrid_projects_the_key_instead_of_asking_for_row_ids():
assert builder._pk_fusion.injected == ["id"]
for leg in (builder._vector_query, builder._fts_query):
assert leg._columns == ["text", "id"]
# A MemWAL table rejects with_row_id before it plans.
# An RTE table rejects with_row_id before it plans.
assert leg._with_row_id is not True
@@ -81,15 +81,15 @@ def test_base_table_hybrid_still_joins_on_row_ids():
assert builder._vector_query._columns is None
def test_wal_hybrid_with_use_memwal_false_joins_on_row_ids():
builder = _wal_hybrid().use_memwal(False)
def test_wal_hybrid_with_use_rte_false_joins_on_row_ids():
builder = _wal_hybrid().use_rte(False)
builder._create_query_builders()
assert builder._pk_fusion is None
def test_wal_hybrid_refuses_an_explicit_with_row_id():
builder = _wal_hybrid().with_row_id(True)
with pytest.raises(NotImplementedError, match="use_memwal"):
with pytest.raises(NotImplementedError, match="use_rte"):
builder._create_query_builders()
@@ -141,7 +141,7 @@ def test_wal_hybrid_refuses_row_id_even_with_return_score_all(row_id_first: bool
else:
builder = builder.rerank(reranker).with_row_id(True)
with pytest.raises(NotImplementedError, match="use_memwal"):
with pytest.raises(NotImplementedError, match="use_rte"):
builder._create_query_builders()
@@ -199,7 +199,7 @@ def test_fallback_explains_the_refusal_to_a_caller_who_asked_for_row_ids():
def run():
raise _REFUSAL
with pytest.raises(NotImplementedError, match="use_memwal") as info:
with pytest.raises(NotImplementedError, match="use_rte") as info:
with_pk_fallback(
run, mock.Mock(), fused=lambda: False, caller_requested_row_id=True
)
python/python/tests/utils.py
+3 -3
View File
@@ -73,10 +73,10 @@ class MockPermutationServer:
},
},
)
if path == f"/v1/table/{self.name}/get_memwal_spec/":
if path == f"/v1/table/{self.name}/get_rte_spec/":
self._read_body(request)
# Null spec: this table has no MemWAL write path.
return self._json(request, {"memwal_spec": None})
# Null spec: this table has no RTE write path.
return self._json(request, {"rte_spec": None})
if path == f"/v1/table/{self.name}/count_rows/":
self._read_body(request)
return self._json(request, self.num_rows)
python/src/lib.rs
+3 -3
View File
@@ -16,8 +16,8 @@ use query::{FTSQuery, HybridQuery, Query, VectorQuery};
use session::Session;
use table::{
AddColumnsResult, AddResult, AlterColumnsResult, DeleteResult, DropColumnsResult, FtsToken,
FunctionErrorFragment, FunctionErrorRecord, FunctionErrors, MemWalWriteSpec, MergeResult,
PyBlobFile, RefreshColumnResult, RefreshMaterializedViewResult, Table,
FunctionErrorFragment, FunctionErrorRecord, FunctionErrors, MergeResult, PyBlobFile,
RTEWriteSpec, RefreshColumnResult, RefreshMaterializedViewResult, Table,
UpdateFieldMetadataResult, UpdateResult,
};
@@ -104,7 +104,7 @@ pub fn _lancedb(py: Python, m: &Bound<'_, PyModule>) -> PyResult<()> {
m.add_class::<UpdateFieldMetadataResult>()?;
m.add_class::<AddResult>()?;
m.add_class::<MergeResult>()?;
m.add_class::<MemWalWriteSpec>()?;
m.add_class::<RTEWriteSpec>()?;
m.add_class::<DeleteResult>()?;
m.add_class::<DropColumnsResult>()?;
m.add_class::<UpdateResult>()?;
python/src/query.rs
+14 -14
View File
@@ -333,7 +333,7 @@ pub struct PyQueryRequest {
pub select_source_columns: Option<HashMap<String, String>>,
pub fast_search: Option<bool>,
pub with_row_id: Option<bool>,
pub use_memwal: Option<bool>,
pub use_rte: Option<bool>,
pub column: Option<String>,
pub query_vector: Option<PyQueryVectors>,
pub nprobes: Option<usize>,
@@ -366,7 +366,7 @@ impl From<AnyQuery> for PyQueryRequest {
select: PySelect(query_request.select),
fast_search: Some(query_request.fast_search),
with_row_id: Some(query_request.with_row_id),
use_memwal: query_request.use_memwal,
use_rte: query_request.use_rte,
column: None,
query_vector: None,
nprobes: None,
@@ -395,7 +395,7 @@ impl From<AnyQuery> for PyQueryRequest {
select: PySelect(vector_query.base.select),
fast_search: Some(vector_query.base.fast_search),
with_row_id: Some(vector_query.base.with_row_id),
use_memwal: vector_query.base.use_memwal,
use_rte: vector_query.base.use_rte,
column: vector_query.column,
query_vector: Some(PyQueryVectors(vector_query.query_vector)),
nprobes: vector_query.nprobes,
@@ -540,8 +540,8 @@ impl Query {
self.inner = self.inner.clone().fast_search();
}
pub fn use_memwal(&mut self, enable: bool) {
self.inner = self.inner.clone().use_memwal(enable);
pub fn use_rte(&mut self, enable: bool) {
self.inner = self.inner.clone().use_rte(enable);
}
pub fn with_row_id(&mut self) {
@@ -706,8 +706,8 @@ impl TakeQuery {
self.inner = self.inner.clone().with_row_id();
}
pub fn use_memwal(&mut self, enable: bool) {
self.inner = self.inner.clone().use_memwal(enable);
pub fn use_rte(&mut self, enable: bool) {
self.inner = self.inner.clone().use_rte(enable);
}
#[pyo3(signature = ())]
@@ -819,8 +819,8 @@ impl FTSQuery {
self.inner = self.inner.clone().fast_search();
}
pub fn use_memwal(&mut self, enable: bool) {
self.inner = self.inner.clone().use_memwal(enable);
pub fn use_rte(&mut self, enable: bool) {
self.inner = self.inner.clone().use_rte(enable);
}
pub fn with_row_id(&mut self) {
@@ -970,8 +970,8 @@ impl VectorQuery {
self.inner = self.inner.clone().fast_search();
}
pub fn use_memwal(&mut self, enable: bool) {
self.inner = self.inner.clone().use_memwal(enable);
pub fn use_rte(&mut self, enable: bool) {
self.inner = self.inner.clone().use_rte(enable);
}
pub fn with_row_id(&mut self) {
@@ -1178,9 +1178,9 @@ impl HybridQuery {
self.inner_fts.postfilter();
}
pub fn use_memwal(&mut self, enable: bool) {
self.inner_vec.use_memwal(enable);
self.inner_fts.use_memwal(enable);
pub fn use_rte(&mut self, enable: bool) {
self.inner_vec.use_rte(enable);
self.inner_fts.use_rte(enable);
}
pub fn add_query_vector(&mut self, vector: Bound<'_, PyAny>) -> PyResult<()> {
python/src/table.rs
+59 -58
View File
@@ -39,13 +39,13 @@ use pyo3::{
mod scannable;
/// Convert `MemWalStats` to a Python dict, preserving the per-table-shard list.
/// Convert `RTEStats` to a Python dict, preserving the per-table-shard list.
///
/// Deliberately not flattened to a table-level summary: a table is N
/// table shards on one node, and the per-shard detail is the reason the
/// endpoint exists — flattening hides the single hot table shard someone opened
/// it to find.
fn memwal_stats_to_py(py: Python<'_>, stats: &lancedb::table::MemWalStats) -> PyResult<Py<PyDict>> {
fn rte_stats_to_py(py: Python<'_>, stats: &lancedb::table::RTEStats) -> PyResult<Py<PyDict>> {
let out = PyDict::new(py);
let table_shards = PyList::empty(py);
for b in &stats.table_shards {
@@ -261,8 +261,8 @@ fn fmt_maintained(maintained: &Option<Vec<String>>) -> String {
}
}
/// Specification selecting Lance's MemWAL LSM-style write path for
/// `merge_insert`.
/// Specification enabling the realtime engine (RTE), which routes
/// `merge_insert` through Lance's MemWAL LSM-style write path.
///
/// Constructed via the `bucket(...)`, `identity(...)`, or `unsharded()`
/// classmethods, then optionally chain `with_maintained_indexes(...)` and
@@ -270,17 +270,17 @@ fn fmt_maintained(maintained: &Option<Vec<String>>) -> String {
/// table has, including ones created later.
#[pyclass(module = "lancedb._lancedb", from_py_object)]
#[derive(Clone, Debug)]
pub struct MemWalWriteSpec {
inner: lancedb::table::MemWalWriteSpec,
pub struct RTEWriteSpec {
inner: lancedb::table::RTEWriteSpec,
}
#[pymethods]
impl MemWalWriteSpec {
impl RTEWriteSpec {
/// Hash-bucket sharding by the unenforced primary key column.
#[staticmethod]
pub fn bucket(column: String, num_buckets: u32) -> Self {
Self {
inner: lancedb::table::MemWalWriteSpec::bucket(column, num_buckets),
inner: lancedb::table::RTEWriteSpec::bucket(column, num_buckets),
}
}
@@ -294,20 +294,20 @@ impl MemWalWriteSpec {
#[staticmethod]
pub fn identity(column: String) -> Self {
Self {
inner: lancedb::table::MemWalWriteSpec::identity(column),
inner: lancedb::table::RTEWriteSpec::identity(column),
}
}
/// No sharding — every `merge_insert` call writes to a single
/// MemWAL shard.
/// RTE shard.
#[staticmethod]
pub fn unsharded() -> Self {
Self {
inner: lancedb::table::MemWalWriteSpec::unsharded(),
inner: lancedb::table::RTEWriteSpec::unsharded(),
}
}
/// Set which indexes the MemWAL maintains. `None` (the default) is
/// Set which indexes the RTE maintains. `None` (the default) is
/// every index the table has, including ones created later; a list is
/// verbatim, and an empty list maintains nothing.
#[pyo3(signature = (indexes))]
@@ -327,33 +327,33 @@ impl MemWalWriteSpec {
pub fn __repr__(&self) -> String {
match &self.inner {
lancedb::table::MemWalWriteSpec::Bucket {
lancedb::table::RTEWriteSpec::Bucket {
column,
num_buckets,
maintained_indexes,
writer_config_defaults,
} => format!(
"MemWalWriteSpec.bucket(column={:?}, num_buckets={}, maintained_indexes={}, writer_config_defaults={:?})",
"RTEWriteSpec.bucket(column={:?}, num_buckets={}, maintained_indexes={}, writer_config_defaults={:?})",
column,
num_buckets,
fmt_maintained(maintained_indexes),
writer_config_defaults,
),
lancedb::table::MemWalWriteSpec::Identity {
lancedb::table::RTEWriteSpec::Identity {
column,
maintained_indexes,
writer_config_defaults,
} => format!(
"MemWalWriteSpec.identity(column={:?}, maintained_indexes={}, writer_config_defaults={:?})",
"RTEWriteSpec.identity(column={:?}, maintained_indexes={}, writer_config_defaults={:?})",
column,
fmt_maintained(maintained_indexes),
writer_config_defaults,
),
lancedb::table::MemWalWriteSpec::Unsharded {
lancedb::table::RTEWriteSpec::Unsharded {
maintained_indexes,
writer_config_defaults,
} => format!(
"MemWalWriteSpec.unsharded(maintained_indexes={}, writer_config_defaults={:?})",
"RTEWriteSpec.unsharded(maintained_indexes={}, writer_config_defaults={:?})",
fmt_maintained(maintained_indexes),
writer_config_defaults,
),
@@ -365,9 +365,9 @@ impl MemWalWriteSpec {
#[getter]
pub fn spec_type(&self) -> &'static str {
match &self.inner {
lancedb::table::MemWalWriteSpec::Bucket { .. } => "bucket",
lancedb::table::MemWalWriteSpec::Identity { .. } => "identity",
lancedb::table::MemWalWriteSpec::Unsharded { .. } => "unsharded",
lancedb::table::RTEWriteSpec::Bucket { .. } => "bucket",
lancedb::table::RTEWriteSpec::Identity { .. } => "identity",
lancedb::table::RTEWriteSpec::Unsharded { .. } => "unsharded",
}
}
@@ -375,9 +375,9 @@ impl MemWalWriteSpec {
#[getter]
pub fn column(&self) -> Option<String> {
match &self.inner {
lancedb::table::MemWalWriteSpec::Bucket { column, .. }
| lancedb::table::MemWalWriteSpec::Identity { column, .. } => Some(column.clone()),
lancedb::table::MemWalWriteSpec::Unsharded { .. } => None,
lancedb::table::RTEWriteSpec::Bucket { column, .. }
| lancedb::table::RTEWriteSpec::Identity { column, .. } => Some(column.clone()),
lancedb::table::RTEWriteSpec::Unsharded { .. } => None,
}
}
@@ -385,12 +385,12 @@ impl MemWalWriteSpec {
#[getter]
pub fn num_buckets(&self) -> Option<u32> {
match &self.inner {
lancedb::table::MemWalWriteSpec::Bucket { num_buckets, .. } => Some(*num_buckets),
lancedb::table::RTEWriteSpec::Bucket { num_buckets, .. } => Some(*num_buckets),
_ => None,
}
}
/// Indexes the MemWAL keeps up to date, or `None` for every supported one.
/// Indexes the RTE keeps up to date, or `None` for every supported one.
#[getter]
pub fn maintained_indexes(&self) -> Option<Vec<String>> {
self.inner.maintained_indexes().map(<[String]>::to_vec)
@@ -403,14 +403,14 @@ impl MemWalWriteSpec {
}
}
impl From<MemWalWriteSpec> for lancedb::table::MemWalWriteSpec {
fn from(spec: MemWalWriteSpec) -> Self {
impl From<RTEWriteSpec> for lancedb::table::RTEWriteSpec {
fn from(spec: RTEWriteSpec) -> Self {
spec.inner
}
}
impl From<lancedb::table::MemWalWriteSpec> for MemWalWriteSpec {
fn from(inner: lancedb::table::MemWalWriteSpec) -> Self {
impl From<lancedb::table::RTEWriteSpec> for RTEWriteSpec {
fn from(inner: lancedb::table::RTEWriteSpec) -> Self {
Self { inner }
}
}
@@ -1636,8 +1636,8 @@ impl Table {
if let Some(use_index) = parameters.use_index {
builder.use_index(use_index);
}
if let Some(use_memwal) = parameters.use_memwal {
builder.use_memwal(use_memwal);
if let Some(use_rte) = parameters.use_rte {
builder.use_rte(use_rte);
}
if let Some(validate_single_shard) = parameters.validate_single_shard {
builder.validate_single_shard(validate_single_shard);
@@ -1662,29 +1662,29 @@ impl Table {
})
}
pub fn enable_memwal<'a>(
pub fn enable_rte<'a>(
self_: PyRef<'a, Self>,
spec: MemWalWriteSpec,
spec: RTEWriteSpec,
) -> PyResult<Bound<'a, PyAny>> {
let inner = self_.inner_ref()?.clone();
let native_spec = lancedb::table::MemWalWriteSpec::from(spec);
let native_spec = lancedb::table::RTEWriteSpec::from(spec);
future_into_py(self_.py(), async move {
inner.enable_memwal(native_spec).await.infer_error()
inner.enable_rte(native_spec).await.infer_error()
})
}
pub fn disable_memwal(self_: PyRef<'_, Self>) -> PyResult<Bound<'_, PyAny>> {
pub fn disable_rte(self_: PyRef<'_, Self>) -> PyResult<Bound<'_, PyAny>> {
let inner = self_.inner_ref()?.clone();
future_into_py(self_.py(), async move {
inner.disable_memwal().await.infer_error()
inner.disable_rte().await.infer_error()
})
}
pub fn get_memwal_spec(self_: PyRef<'_, Self>) -> PyResult<Bound<'_, PyAny>> {
pub fn get_rte_spec(self_: PyRef<'_, Self>) -> PyResult<Bound<'_, PyAny>> {
let inner = self_.inner_ref()?.clone();
future_into_py(self_.py(), async move {
let spec = inner.get_memwal_spec().await.infer_error()?;
Ok(spec.map(MemWalWriteSpec::from))
let spec = inner.get_rte_spec().await.infer_error()?;
Ok(spec.map(RTEWriteSpec::from))
})
}
@@ -1703,54 +1703,55 @@ impl Table {
Ok(())
}
/// Converge the table's MemWAL write path into its base table.
/// Converge the table's RTE write path into its base table.
///
/// Best-effort: with writes flowing, new rows may land after the last
/// pass. Errors if the table stops making progress.
pub fn checkpoint_memwal(self_: PyRef<'_, Self>) -> PyResult<Bound<'_, PyAny>> {
pub fn checkpoint_rte(self_: PyRef<'_, Self>) -> PyResult<Bound<'_, PyAny>> {
let inner = self_.inner_ref()?.clone();
future_into_py(self_.py(), async move {
inner.checkpoint_memwal().await.infer_error()
inner.checkpoint_rte().await.infer_error()
})
}
/// Freeze every table shard's active memtable into an SSTable.
pub fn flush_memwal(self_: PyRef<'_, Self>) -> PyResult<Bound<'_, PyAny>> {
pub fn flush_rte(self_: PyRef<'_, Self>) -> PyResult<Bound<'_, PyAny>> {
let inner = self_.inner_ref()?.clone();
future_into_py(self_.py(), async move {
inner.flush_memwal().await.infer_error()
})
future_into_py(
self_.py(),
async move { inner.flush_rte().await.infer_error() },
)
}
/// Trigger a background SSTable compaction pass per table shard. Returns once the
/// passes are dispatched, not once they finish — watch `get_memwal_stats`.
pub fn compact_memwal(self_: PyRef<'_, Self>) -> PyResult<Bound<'_, PyAny>> {
/// passes are dispatched, not once they finish — watch `get_rte_stats`.
pub fn compact_rte(self_: PyRef<'_, Self>) -> PyResult<Bound<'_, PyAny>> {
let inner = self_.inner_ref()?.clone();
future_into_py(self_.py(), async move {
inner.compact_memwal().await.infer_error()
inner.compact_rte().await.infer_error()
})
}
/// Live MemWAL state, or `None` when the MemWAL write path is not enabled.
/// Live RTE state, or `None` when the RTE write path is not enabled.
#[pyo3(signature = (include_sstable_rows=false))]
pub fn get_memwal_stats(
pub fn get_rte_stats(
self_: PyRef<'_, Self>,
include_sstable_rows: bool,
) -> PyResult<Bound<'_, PyAny>> {
let inner = self_.inner_ref()?.clone();
future_into_py(self_.py(), async move {
let stats = inner
.get_memwal_stats(include_sstable_rows)
.get_rte_stats(include_sstable_rows)
.await
.infer_error()?;
Python::attach(|py| stats.map(|s| memwal_stats_to_py(py, &s)).transpose())
Python::attach(|py| stats.map(|s| rte_stats_to_py(py, &s)).transpose())
})
}
pub fn close_memwal_writers(self_: PyRef<'_, Self>) -> PyResult<Bound<'_, PyAny>> {
pub fn close_rte_writers(self_: PyRef<'_, Self>) -> PyResult<Bound<'_, PyAny>> {
let inner = self_.inner_ref()?.clone();
future_into_py(self_.py(), async move {
inner.close_memwal_writers().await.infer_error()
inner.close_rte_writers().await.infer_error()
})
}
@@ -2082,7 +2083,7 @@ pub struct MergeInsertParams {
when_not_matched_by_source_condition_expr: Option<PyExpr>,
timeout: Option<std::time::Duration>,
use_index: Option<bool>,
use_memwal: Option<bool>,
use_rte: Option<bool>,
validate_single_shard: Option<bool>,
}
rust/lancedb/src/database/listing.rs
+3 -3
View File
@@ -540,7 +540,7 @@ impl ListingDatabase {
// pairs survive (or none existed in the first place), so an empty
// string here must be treated the same as "no query" — otherwise
// every table URI ends up with a trailing `?`, which makes downstream
// sub-paths (e.g. MemWAL gen paths) re-parse as path=<base table> +
// sub-paths (e.g. RTE gen paths) re-parse as path=<base table> +
// query=<sub-path>, causing Lance to find the base table dataset
// when looking up the sub-path.
let query_string = url.query().filter(|q| !q.is_empty()).map(|s| s.to_string());
@@ -3104,7 +3104,7 @@ mod tests {
///
/// Earlier, `query_pairs_mut().clear()` left the URL with
/// `query=Some("")`, which then propagated as a trailing `?` on
/// every table URI. Sub-path lookups against that URI (e.g. MemWAL
/// every table URI. Sub-path lookups against that URI (e.g. RTE
/// `<table_uri>/_mem_wal/<shard>/<rand>_gen_<n>`) re-parsed as
/// `path=<base table>` + `query=/_mem_wal/...`, causing
/// `Dataset::write` to find the base table dataset and falsely
@@ -3178,7 +3178,7 @@ mod tests {
/// `url::Url::parse` and the `query_pairs_mut()` path) must not
/// append a trailing `?` to per-table URIs when the input URI has
/// no query string. Sub-path lookups against such a URI (e.g.
/// MemWAL `<table_uri>/_mem_wal/<shard>/<rand>_gen_<n>`) re-parse
/// RTE `<table_uri>/_mem_wal/<shard>/<rand>_gen_<n>`) re-parse
/// as `path=<base table>` + `query=/_mem_wal/...`, causing
/// `Dataset::write` to find the base table dataset and falsely
/// report `Dataset already exists`.
rust/lancedb/src/dataloader/permutation/builder.rs
+10 -13
View File
@@ -263,17 +263,17 @@ impl PermutationBuilder {
}
// Unflushed rows have no row id, so a permutation cannot address them.
match self.base_table.base_table().get_memwal_spec().await {
match self.base_table.base_table().get_rte_spec().await {
Ok(Some(_)) => {
return Err(Error::NotSupported {
message: "the data loader does not support tables with a MemWAL write \
message: "the data loader does not support tables with an RTE write \
spec: rows that have not been flushed to the base table \
have no row id, so a permutation cannot reference them"
.to_string(),
});
}
Ok(None) => {}
// No MemWAL write path means no spec.
// No RTE write path means no spec.
Err(Error::NotSupported { .. }) => {}
Err(err) => return Err(err),
}
@@ -502,9 +502,9 @@ mod tests {
)
.unwrap()
}
"/v1/table/remote_base/get_memwal_spec/" => http::Response::builder()
"/v1/table/remote_base/get_rte_spec/" => http::Response::builder()
.status(200)
.body(br#"{"memwal_spec":null}"#.to_vec())
.body(br#"{"rte_spec":null}"#.to_vec())
.unwrap(),
"/v1/table/remote_base/count_rows/" => {
let body = body.unwrap();
@@ -750,12 +750,12 @@ mod tests {
/// permutation cannot reference them. Reading the base table alone would drop
/// them from training without saying so, so the table is refused instead.
#[tokio::test]
async fn test_permutation_rejects_memwal_spec() {
use crate::table::MemWalWriteSpec;
async fn test_permutation_rejects_rte_spec() {
use crate::table::RTEWriteSpec;
use arrow_array::{Int32Array, RecordBatchIterator};
use arrow_schema::{DataType, Field, Schema};
// MemWAL needs a real dataset directory and a non-nullable primary key.
// RTE needs a real dataset directory and a non-nullable primary key.
let temp_dir = tempfile::tempdir().unwrap();
let db = connect(temp_dir.path().to_str().unwrap())
.execute()
@@ -778,14 +778,11 @@ mod tests {
.unwrap();
table.set_unenforced_primary_key(["idx"]).await.unwrap();
table
.enable_memwal(MemWalWriteSpec::unsharded())
.await
.unwrap();
table.enable_rte(RTEWriteSpec::unsharded()).await.unwrap();
let err = PermutationBuilder::new(table).build().await.unwrap_err();
assert!(
err.to_string().contains("MemWAL write spec"),
err.to_string().contains("RTE write spec"),
"expected the pre-check to refuse the table, got: {err}"
);
}
rust/lancedb/src/lib.rs
+1 -1
View File
@@ -167,7 +167,7 @@
//! # }
//! ```
// The MemWAL read path (`table::query::lsm`) deepens the `create_plan` future's
// The RTE read path (`table::query::lsm`) deepens the `create_plan` future's
// type graph enough to overflow the default trait-recursion limit while evaluating
// auto-traits (`Send`) through the Linux io_uring build's moka cache. Raise it.
#![recursion_limit = "256"]
rust/lancedb/src/materialized_view/refresh.rs
+2 -4
View File
@@ -288,15 +288,13 @@ fn refresh_conflict_filter(dataset: &Dataset) -> Result<KeyExistenceFilter> {
Ok(filter.build())
}
/// Reject MemWAL/MemWAL state because an ordinary dataset scan cannot see its
/// Reject RTE/RTE state because an ordinary dataset scan cannot see its
/// un-compacted tiers.
pub(crate) async fn ensure_no_mem_wal(dataset: &Dataset, role: &str, name: &str) -> Result<()> {
let retained = !dataset.list_mem_wal_latest_shard_ids().await?.is_empty();
if retained || dataset.mem_wal_index_details().await?.is_some() {
return Err(Error::NotSupported {
message: format!(
"{role} '{name}' has a MemWAL write spec or retained un-compacted rows"
),
message: format!("{role} '{name}' has an RTE write spec or retained un-compacted rows"),
});
}
Ok(())
rust/lancedb/src/query.rs
+19 -19
View File
@@ -544,25 +544,25 @@ pub trait QueryBase {
/// This allows ordering query results by one or more columns in either ascending or descending order.
fn order_by(self, ordering: Option<Vec<ColumnOrdering>>) -> Self;
/// Control MemWAL read routing for this query.
/// Control RTE read routing for this query.
///
/// By default (unset), when the table carries a MemWAL write spec (see
/// [`crate::Table::enable_memwal`]), reads are routed through the LSM
/// By default (unset), when the table carries an RTE write spec (see
/// [`crate::Table::enable_rte`]), reads are routed through the LSM
/// scanner so they also return data written via the `merge_insert` LSM path
/// that has not yet been compacted into the base table (active/frozen
/// memtables and SSTables); a table without a spec reads the base
/// table.
///
/// - `use_memwal(true)` forces MemWAL routing and errors if the table has no
/// MemWAL write spec.
/// - `use_memwal(false)` bypasses the MemWAL and reads the base table only,
/// - `use_rte(true)` forces RTE routing and errors if the table has no
/// RTE write spec.
/// - `use_rte(false)` bypasses the RTE and reads the base table only,
/// even when a spec is present.
///
/// Note: the LSM scanner does not support every query shape (e.g. reranking,
/// hybrid search, `order_by`). On a MemWAL table those shapes error unless
/// `use_memwal(false)` is set, because a base-only read would silently
/// exclude un-compacted MemWAL data.
fn use_memwal(self, enable: bool) -> Self;
/// hybrid search, `order_by`). On an RTE table those shapes error unless
/// `use_rte(false)` is set, because a base-only read would silently
/// exclude un-compacted RTE data.
fn use_rte(self, enable: bool) -> Self;
}
pub trait HasQuery {
@@ -634,8 +634,8 @@ impl<T: HasQuery> QueryBase for T {
self
}
fn use_memwal(mut self, enable: bool) -> Self {
self.mut_query().use_memwal = Some(enable);
fn use_rte(mut self, enable: bool) -> Self {
self.mut_query().use_rte = Some(enable);
self
}
}
@@ -898,19 +898,19 @@ pub struct QueryRequest {
/// This allows ordering query results by one or more columns in either ascending or descending order.
pub order_by: Option<Vec<ColumnOrdering>>,
/// Controls MemWAL read routing. When unset (the default), a query against a
/// table that carries a MemWAL write spec (see
/// [`crate::Table::enable_memwal`]) is routed through the LSM scanner so
/// Controls RTE read routing. When unset (the default), a query against a
/// table that carries an RTE write spec (see
/// [`crate::Table::enable_rte`]) is routed through the LSM scanner so
/// it also sees data written via the `merge_insert` LSM path that has not yet
/// been compacted into the base table — the active and frozen in-memory
/// memtables and the SSTables, deduplicated by primary key
/// against the base table (newest generation wins); a table without a spec
/// reads the base table.
///
/// - `Some(true)` forces MemWAL routing and errors if the table has no MemWAL
/// - `Some(true)` forces RTE routing and errors if the table has no RTE
/// write spec.
/// - `Some(false)` reads only the base table, bypassing the MemWAL.
pub use_memwal: Option<bool>,
/// - `Some(false)` reads only the base table, bypassing the RTE.
pub use_rte: Option<bool>,
}
impl Default for QueryRequest {
@@ -930,7 +930,7 @@ impl Default for QueryRequest {
norm: None,
disable_scoring_autoprojection: false,
order_by: None,
use_memwal: None,
use_rte: None,
}
}
}
rust/lancedb/src/query/wal_fusion.rs
+17 -17
View File
@@ -1,14 +1,14 @@
// SPDX-License-Identifier: Apache-2.0
// SPDX-FileCopyrightText: Copyright The LanceDB Authors
//! WAL-PK-FUSION: hybrid search's primary-key fallback for MemWAL tables.
//! WAL-PK-FUSION: hybrid search's primary-key fallback for RTE tables.
//!
//! Hybrid search joins its vector and full-text legs on `_rowid`, which a
//! MemWAL table cannot supply: the fresh tier has no stable row id, so the
//! RTE table cannot supply: the fresh tier has no stable row id, so the
//! server rejects `with_row_id` before it plans. When that happens the legs are
//! joined on a surrogate `_rowid` built from the table's primary key instead.
//!
//! This is a stopgap until MemWAL supports `_rowid`, and everything it needs
//! This is a stopgap until RTE supports `_rowid`, and everything it needs
//! lives here so it can be removed in one pass:
//!
//! 1. Delete this file and its `mod wal_fusion;` line in `query.rs`.
@@ -31,44 +31,44 @@ use lance_core::utils::parse::str_is_truthy;
use super::{Select, VectorQuery};
use crate::error::{Error, Result};
/// The part of the MemWAL `with_row_id` refusal that identifies it.
/// The part of the RTE `with_row_id` refusal that identifies it.
///
/// Both layers raise this same sentence verbatim — sophon's
/// `LSM_WITH_ROW_ID_UNSUPPORTED` says so in as many words, and the OSS LSM
/// scanner builds it from the same words — so this substring survives either
/// path and the wrapping each adds. It is a coupling to a message rather than
/// a code, because the refusal arrives as a generic 400; if the server ever
/// rewords it the fallback stops firing and hybrid on a MemWAL table goes back
/// rewords it the fallback stops firing and hybrid on an RTE table goes back
/// to erroring visibly, which is loud rather than silently wrong.
const WAL_ROW_ID_REFUSAL: &str = "does not support with_row_id";
/// Whether `error` is the server declining `_rowid` because the table is
/// MemWAL-backed, as opposed to any other failure the query might hit.
/// RTE-backed, as opposed to any other failure the query might hit.
fn is_wal_row_id_refusal(error: &Error) -> bool {
error.to_string().contains(WAL_ROW_ID_REFUSAL)
}
/// Why a hybrid query on a MemWAL table cannot hand back `_rowid`, raised
/// Why a hybrid query on an RTE table cannot hand back `_rowid`, raised
/// whether the refusal was anticipated or came back from the server.
fn wal_row_id_unsupported() -> Error {
Error::NotSupported {
message: "hybrid search on a MemWAL table cannot return _rowid: the fresh tier has no \
message: "hybrid search on an RTE table cannot return _rowid: the fresh tier has no \
stable row id, and the ids the fusion joins on are synthesized from the \
primary key. Set use_memwal(false) to read the base table only (results will \
exclude un-compacted MemWAL data)"
primary key. Set use_rte(false) to read the base table only (results will \
exclude un-compacted RTE data)"
.to_string(),
}
}
/// Run a hybrid query, joining its legs on the primary key instead of `_rowid`
/// when the table is MemWAL-backed.
/// when the table is RTE-backed.
///
/// `run(None)` must join the legs on `_rowid`; `run(Some(fusion))` must build
/// them with `fusion` instead.
///
/// `_rowid` is asked for first and the primary key used only when the server
/// refuses, rather than paying a round trip up front to find out. Only a
/// MemWAL table refuses, so probing would tax every table to learn something
/// RTE table refuses, so probing would tax every table to learn something
/// almost none of them need — and the refusal has to be handled regardless,
/// since a write spec installed elsewhere can arrive between any two queries.
/// The refusal costs nothing server-side: it is raised before the query is
@@ -80,7 +80,7 @@ where
{
// A previous query on this table was already refused, so skip straight to
// the key.
if query.request.base.use_memwal != Some(false) && query.parent.hybrid_pk_fusion_learned() {
if query.request.base.use_rte != Some(false) && query.parent.hybrid_pk_fusion_learned() {
return run(Some(PkFusion::resolve(query).await?)).await;
}
match run(None).await {
@@ -130,7 +130,7 @@ impl PkFusionMemory {
}
}
/// How the legs of a hybrid query on a MemWAL table are joined: on a surrogate
/// How the legs of a hybrid query on an RTE table are joined: on a surrogate
/// `_rowid` derived client-side from the primary key.
pub(super) struct PkFusion {
columns: Vec<String>,
@@ -152,7 +152,7 @@ impl PkFusion {
Ok(Self { columns, injected })
}
/// Make a leg return the key columns, in place of the `_rowid` a MemWAL
/// Make a leg return the key columns, in place of the `_rowid` an RTE
/// table would refuse.
pub(super) fn prepare_leg(&self, select: &mut Select) {
inject_pk_columns(select, &self.columns);
@@ -266,7 +266,7 @@ fn pk_columns(schema: &arrow_schema::Schema) -> Vec<String> {
/// Replace each batch's join column with a dense surrogate derived from
/// `pk_columns`, in place and in the order given.
///
/// The fusion joins its two legs on [`ROW_ID`], which a MemWAL table cannot
/// The fusion joins its two legs on [`ROW_ID`], which an RTE table cannot
/// supply — the fresh tier has no stable row id. It only ever compares the
/// column for equality, though (`merge_results` dedups, the rerankers group and
/// the score restore looks up), so any injective mapping of the primary key
@@ -280,7 +280,7 @@ fn pk_columns(schema: &arrow_schema::Schema) -> Vec<String> {
fn stamp_surrogate_row_ids(batches: &mut [RecordBatch], pk_columns: &[String]) -> Result<()> {
if pk_columns.is_empty() {
return Err(Error::InvalidInput {
message: "hybrid search on a MemWAL table needs an unenforced primary key to \
message: "hybrid search on an RTE table needs an unenforced primary key to \
deduplicate on, and this table declares none"
.to_string(),
});
rust/lancedb/src/remote/table.rs
+124 -125
View File
@@ -30,15 +30,15 @@ use crate::table::BranchDiff;
use crate::table::CherryPickResult;
use crate::table::DeleteResult;
use crate::table::DropColumnsResult;
use crate::table::MemWalStats;
use crate::table::MemWalWriteSpec;
use crate::table::MergeResult;
use crate::table::RTEStats;
use crate::table::RTEWriteSpec;
use crate::table::Tags;
use crate::table::UpdateResult;
use crate::table::memwal_stats::GetMemWalStatsResponse;
use crate::table::merge::MergeFilter;
use crate::table::primary_key;
use crate::table::query::create_multi_vector_plan;
use crate::table::rte_stats::GetRTEStatsResponse;
use crate::table::write_progress::FinishOnDrop;
use crate::table::{
AlterColumnsResult, FieldMetadataUpdate, RefreshColumnResult, UpdateFieldMetadataResult,
@@ -1020,10 +1020,10 @@ impl<S: HttpSend> RemoteTable<S> {
) -> Result<()> {
params.check_filter()?;
body["prefilter"] = params.prefilter.into();
// Only forward use_memwal when explicitly set; a server that predates it
// Only forward use_rte when explicitly set; a server that predates it
// ignores the field and routes as it would by default.
if let Some(use_memwal) = params.use_memwal {
body["use_memwal"] = serde_json::Value::Bool(use_memwal);
if let Some(use_rte) = params.use_rte {
body["use_rte"] = serde_json::Value::Bool(use_rte);
}
if let Some(offset) = params.offset {
body["offset"] = serde_json::Value::Number(serde_json::Number::from(offset));
@@ -3236,43 +3236,42 @@ impl<S: HttpSend> BaseTable for RemoteTable<S> {
Ok(())
}
async fn flush_memwal(&self) -> Result<()> {
async fn flush_rte(&self) -> Result<()> {
let request = self
.client
.post(&format!("/v1/table/{}/flush_memwal/", self.identifier));
.post(&format!("/v1/table/{}/flush_rte/", self.identifier));
self.send_lsm_route(request).await?;
Ok(())
}
async fn compact_memwal(&self) -> Result<()> {
async fn compact_rte(&self) -> Result<()> {
let request = self
.client
.post(&format!("/v1/table/{}/compact_memwal/", self.identifier));
.post(&format!("/v1/table/{}/compact_rte/", self.identifier));
self.send_lsm_route(request).await?;
Ok(())
}
async fn get_memwal_stats(&self, include_sstable_rows: bool) -> Result<Option<MemWalStats>> {
// Read-semantics POST, like `get_memwal_spec`.
async fn get_rte_stats(&self, include_sstable_rows: bool) -> Result<Option<RTEStats>> {
// Read-semantics POST, like `get_rte_spec`.
let request = self
.client
.post(&format!("/v1/table/{}/get_memwal_stats/", self.identifier))
.post(&format!("/v1/table/{}/get_rte_stats/", self.identifier))
.json(&serde_json::json!({
"include_sstable_rows": include_sstable_rows,
}));
let (request_id, response) = self.send_lsm_route(request).await?;
let body = response.text().await.err_to_http(request_id.clone())?;
let parsed: GetMemWalStatsResponse =
serde_json::from_str(&body).map_err(|e| Error::Http {
source: format!("Failed to parse get_memwal_stats response: {e}").into(),
request_id,
status_code: None,
})?;
// `null` — and only — when the table has no MemWAL write path.
Ok(parsed.memwal_stats)
let parsed: GetRTEStatsResponse = serde_json::from_str(&body).map_err(|e| Error::Http {
source: format!("Failed to parse get_rte_stats response: {e}").into(),
request_id,
status_code: None,
})?;
// `null` — and only — when the table has no RTE write path.
Ok(parsed.rte_stats)
}
async fn enable_memwal(&self, spec: MemWalWriteSpec) -> Result<()> {
async fn enable_rte(&self, spec: RTEWriteSpec) -> Result<()> {
self.check_mutable().await?;
// Map the spec onto the server's request DTO. `sharding` is internally
@@ -3280,7 +3279,7 @@ impl<S: HttpSend> BaseTable for RemoteTable<S> {
// `maintained_indexes` asks the server to resolve every maintainable
// index at HEAD; a list is verbatim, an empty one meaning none.
let sharding = match &spec {
MemWalWriteSpec::Bucket {
RTEWriteSpec::Bucket {
column,
num_buckets,
..
@@ -3289,11 +3288,11 @@ impl<S: HttpSend> BaseTable for RemoteTable<S> {
"column": column,
"num_buckets": num_buckets,
}),
MemWalWriteSpec::Identity { column, .. } => serde_json::json!({
RTEWriteSpec::Identity { column, .. } => serde_json::json!({
"mode": "identity",
"column": column,
}),
MemWalWriteSpec::Unsharded { .. } => serde_json::json!({ "mode": "unsharded" }),
RTEWriteSpec::Unsharded { .. } => serde_json::json!({ "mode": "unsharded" }),
};
let body = serde_json::json!({
"sharding": sharding,
@@ -3303,40 +3302,40 @@ impl<S: HttpSend> BaseTable for RemoteTable<S> {
let request = self
.client
.post(&format!("/v1/table/{}/enable_memwal/", self.identifier))
.post(&format!("/v1/table/{}/enable_rte/", self.identifier))
.json(&body);
let (request_id, response) = self.send(request, true).await?;
self.check_table_response(&request_id, response).await?;
Ok(())
}
async fn disable_memwal(&self) -> Result<()> {
async fn disable_rte(&self) -> Result<()> {
self.check_mutable().await?;
let request = self
.client
.post(&format!("/v1/table/{}/disable_memwal/", self.identifier));
.post(&format!("/v1/table/{}/disable_rte/", self.identifier));
let (request_id, response) = self.send(request, true).await?;
self.check_table_response(&request_id, response).await?;
self.wal_pk_fusion.forget(); // WAL-PK-FUSION: delete.
Ok(())
}
async fn get_memwal_spec(&self) -> Result<Option<MemWalWriteSpec>> {
async fn get_rte_spec(&self) -> Result<Option<RTEWriteSpec>> {
// Read counterpart to set/unset, resolved server-side against HEAD. The
// server reads the spec from the `__lance_mem_wal` system index (shard
// column mapped from its Lance field id against the current schema) and
// re-encodes it into the same sophon-owned shape the set endpoint
// accepts — no lance/lancedb types cross the wire. `memwal_spec` is
// null when the MemWAL write path is not enabled for the table.
// accepts — no lance/lancedb types cross the wire. `rte_spec` is
// null when the RTE write path is not enabled for the table.
let request = self
.client
.post(&format!("/v1/table/{}/get_memwal_spec/", self.identifier));
.post(&format!("/v1/table/{}/get_rte_spec/", self.identifier));
let (request_id, response) = self.send(request, true).await?;
let response = self.check_table_response(&request_id, response).await?;
let body = response.text().await.err_to_http(request_id.clone())?;
// Mirror of sophon's `Sharding` (internally tagged on `mode`) and
// `MemWalWriteSpecBody` / `GetMemWalSpecResponse`.
// `RTEWriteSpecBody` / `GetRTESpecResponse`.
#[derive(Deserialize)]
#[serde(tag = "mode", rename_all = "snake_case")]
enum Sharding {
@@ -3345,7 +3344,7 @@ impl<S: HttpSend> BaseTable for RemoteTable<S> {
Identity { column: String },
}
#[derive(Deserialize)]
struct MemWalWriteSpecBody {
struct RTEWriteSpecBody {
sharding: Sharding,
/// `null` selects every index the table has; `[]` selects none.
#[serde(default)]
@@ -3354,19 +3353,18 @@ impl<S: HttpSend> BaseTable for RemoteTable<S> {
writer_config_defaults: std::collections::HashMap<String, String>,
}
#[derive(Deserialize)]
struct GetMemWalSpecResponse {
memwal_spec: Option<MemWalWriteSpecBody>,
struct GetRTESpecResponse {
rte_spec: Option<RTEWriteSpecBody>,
}
let parsed: GetMemWalSpecResponse =
serde_json::from_str(&body).map_err(|e| Error::Http {
source: format!("Failed to parse get_memwal_spec response: {}", e).into(),
request_id,
status_code: None,
})?;
let parsed: GetRTESpecResponse = serde_json::from_str(&body).map_err(|e| Error::Http {
source: format!("Failed to parse get_rte_spec response: {}", e).into(),
request_id,
status_code: None,
})?;
let Some(body) = parsed.memwal_spec else {
// The MemWAL write path is not enabled for this table.
let Some(body) = parsed.rte_spec else {
// The RTE write path is not enabled for this table.
return Ok(None);
};
@@ -3374,9 +3372,9 @@ impl<S: HttpSend> BaseTable for RemoteTable<S> {
Sharding::Bucket {
column,
num_buckets,
} => MemWalWriteSpec::bucket(column, num_buckets),
Sharding::Identity { column } => MemWalWriteSpec::identity(column),
Sharding::Unsharded => MemWalWriteSpec::unsharded(),
} => RTEWriteSpec::bucket(column, num_buckets),
Sharding::Identity { column } => RTEWriteSpec::identity(column),
Sharding::Unsharded => RTEWriteSpec::unsharded(),
}
.with_maintained_indexes(body.maintained_indexes)
.with_writer_config_defaults(body.writer_config_defaults);
@@ -3979,10 +3977,10 @@ pub struct MergeInsertRequest {
// (the default is true)
#[serde(skip_serializing_if = "is_true")]
use_index: bool,
// Only serialize use_memwal when explicitly set (Some); a server that predates
// Only serialize use_rte when explicitly set (Some); a server that predates
// it ignores the field and routes as it would by default.
#[serde(skip_serializing_if = "Option::is_none")]
use_memwal: Option<bool>,
use_rte: Option<bool>,
}
impl MergeInsertRequest {
@@ -4040,7 +4038,7 @@ impl TryFrom<MergeInsertBuilder> for MergeInsertRequest {
when_not_matched_by_source_delete_filt,
// Only serialize use_index when it's false for backwards compatibility
use_index: value.use_index,
use_memwal: value.use_memwal,
use_rte: value.use_rte,
})
}
}
@@ -10296,10 +10294,10 @@ mod tests {
}
#[tokio::test]
async fn test_enable_memwal_unsharded() {
async fn test_enable_rte_unsharded() {
let table = Table::new_with_handler("my_table", |request| {
assert_eq!(request.method(), "POST");
assert_eq!(request.url().path(), "/v1/table/my_table/enable_memwal/");
assert_eq!(request.url().path(), "/v1/table/my_table/enable_rte/");
let body = request.body().unwrap().as_bytes().unwrap();
let body: serde_json::Value = serde_json::from_slice(body).unwrap();
assert_eq!(body["sharding"], serde_json::json!({ "mode": "unsharded" }));
@@ -10313,17 +10311,17 @@ mod tests {
.body(r#"{"maintained_indexes":["id_idx"]}"#)
.unwrap()
});
let spec = crate::table::MemWalWriteSpec::unsharded()
let spec = crate::table::RTEWriteSpec::unsharded()
.with_maintained_indexes(vec!["id_idx".to_string()])
.with_writer_config_defaults([("max_memtable_rows", "1000")]);
table.enable_memwal(spec).await.unwrap();
table.enable_rte(spec).await.unwrap();
}
#[tokio::test]
async fn test_enable_memwal_bucket() {
async fn test_enable_rte_bucket() {
let table = Table::new_with_handler("my_table", |request| {
assert_eq!(request.method(), "POST");
assert_eq!(request.url().path(), "/v1/table/my_table/enable_memwal/");
assert_eq!(request.url().path(), "/v1/table/my_table/enable_rte/");
let body = request.body().unwrap().as_bytes().unwrap();
let body: serde_json::Value = serde_json::from_slice(body).unwrap();
assert_eq!(
@@ -10335,14 +10333,14 @@ mod tests {
http::Response::builder().status(200).body("{}").unwrap()
});
table
.enable_memwal(crate::table::MemWalWriteSpec::bucket("id", 16))
.enable_rte(crate::table::RTEWriteSpec::bucket("id", 16))
.await
.unwrap();
}
/// `[]` (none) must stay distinguishable on the wire from null (all).
#[tokio::test]
async fn test_enable_memwal_no_maintained_indexes() {
async fn test_enable_rte_no_maintained_indexes() {
let table = Table::new_with_handler("my_table", |request| {
let body = request.body().unwrap().as_bytes().unwrap();
let body: serde_json::Value = serde_json::from_slice(body).unwrap();
@@ -10350,18 +10348,18 @@ mod tests {
http::Response::builder().status(200).body("{}").unwrap()
});
table
.enable_memwal(
crate::table::MemWalWriteSpec::bucket("id", 16).with_maintained_indexes(Vec::new()),
.enable_rte(
crate::table::RTEWriteSpec::bucket("id", 16).with_maintained_indexes(Vec::new()),
)
.await
.unwrap();
}
#[tokio::test]
async fn test_enable_memwal_identity() {
async fn test_enable_rte_identity() {
let table = Table::new_with_handler("my_table", |request| {
assert_eq!(request.method(), "POST");
assert_eq!(request.url().path(), "/v1/table/my_table/enable_memwal/");
assert_eq!(request.url().path(), "/v1/table/my_table/enable_rte/");
let body = request.body().unwrap().as_bytes().unwrap();
let body: serde_json::Value = serde_json::from_slice(body).unwrap();
assert_eq!(
@@ -10371,32 +10369,32 @@ mod tests {
http::Response::builder().status(200).body("{}").unwrap()
});
table
.enable_memwal(crate::table::MemWalWriteSpec::identity("tenant"))
.enable_rte(crate::table::RTEWriteSpec::identity("tenant"))
.await
.unwrap();
}
#[tokio::test]
async fn test_disable_memwal() {
async fn test_disable_rte() {
let table = Table::new_with_handler("my_table", |request| {
assert_eq!(request.method(), "POST");
assert_eq!(request.url().path(), "/v1/table/my_table/disable_memwal/");
assert_eq!(request.url().path(), "/v1/table/my_table/disable_rte/");
http::Response::builder().status(200).body("{}").unwrap()
});
table.disable_memwal().await.unwrap();
table.disable_rte().await.unwrap();
}
#[tokio::test]
async fn test_get_memwal_spec() {
async fn test_get_rte_spec() {
let table = Table::new_with_handler("my_table", |request| {
assert_eq!(request.method(), "POST");
assert_eq!(request.url().path(), "/v1/table/my_table/get_memwal_spec/");
assert_eq!(request.url().path(), "/v1/table/my_table/get_rte_spec/");
// The server resolves the spec and re-encodes it into the same
// sophon-owned shape the set endpoint accepts (`Sharding` internally
// tagged on `mode`, wrapped in `memwal_spec`).
// tagged on `mode`, wrapped in `rte_spec`).
let response = serde_json::json!({
"memwal_spec": {
"rte_spec": {
"sharding": { "mode": "bucket", "column": "id", "num_buckets": 4 },
"maintained_indexes": ["id_idx"],
"writer_config_defaults": { "durable_write": "false" },
@@ -10409,12 +10407,12 @@ mod tests {
});
let spec = table
.get_memwal_spec()
.get_rte_spec()
.await
.unwrap()
.expect("a spec should be reported");
match spec {
crate::table::MemWalWriteSpec::Bucket {
crate::table::RTEWriteSpec::Bucket {
column,
num_buckets,
maintained_indexes,
@@ -10441,13 +10439,13 @@ mod tests {
#[case::no_index(serde_json::json!([]), Some(vec![]))]
#[case::named(serde_json::json!(["id_idx"]), Some(vec!["id_idx".to_string()]))]
#[tokio::test]
async fn test_get_memwal_spec_round_trips_the_selection(
async fn test_get_rte_spec_round_trips_the_selection(
#[case] reported: serde_json::Value,
#[case] expected: Option<Vec<String>>,
) {
let table = Table::new_with_handler("my_table", move |_| {
let response = serde_json::json!({
"memwal_spec": {
"rte_spec": {
"sharding": { "mode": "unsharded" },
"maintained_indexes": reported,
"writer_config_defaults": {},
@@ -10460,7 +10458,7 @@ mod tests {
});
let spec = table
.get_memwal_spec()
.get_rte_spec()
.await
.unwrap()
.expect("a spec should be reported");
@@ -10472,23 +10470,23 @@ mod tests {
}
#[tokio::test]
async fn test_get_memwal_spec_absent() {
async fn test_get_rte_spec_absent() {
let table = Table::new_with_handler("my_table", |request| {
assert_eq!(request.url().path(), "/v1/table/my_table/get_memwal_spec/");
// Null spec → the MemWAL write path is not enabled.
let response = serde_json::json!({ "memwal_spec": null });
assert_eq!(request.url().path(), "/v1/table/my_table/get_rte_spec/");
// Null spec → the RTE write path is not enabled.
let response = serde_json::json!({ "rte_spec": null });
http::Response::builder()
.status(200)
.body(response.to_string())
.unwrap()
});
assert!(table.get_memwal_spec().await.unwrap().is_none());
assert!(table.get_rte_spec().await.unwrap().is_none());
}
/// Build a `get_memwal_stats` body for one table shard holding `sstables`.
/// Build a `get_rte_stats` body for one table shard holding `sstables`.
fn stats_body(sstables: &[u64], compacting: bool) -> String {
serde_json::json!({
"memwal_stats": {
"rte_stats": {
"table_shards": [{
"shard_id": "b0",
"status": "Active",
@@ -10508,7 +10506,7 @@ mod tests {
.to_string()
}
/// `flush_memwal` / `compact_memwal` answer 202 with no body at all.
/// `flush_rte` / `compact_rte` answer 202 with no body at all.
fn accepted() -> http::Response<String> {
http::Response::builder()
.status(202)
@@ -10529,18 +10527,18 @@ mod tests {
let seen = compacts.clone();
let table = Table::new_with_handler("my_table", move |request| {
let path = request.url().path().to_string();
if path.contains("compact_memwal") {
if path.contains("compact_rte") {
seen.fetch_add(1, std::sync::atomic::Ordering::SeqCst);
panic!("an already-converged table must issue no compact calls");
}
if path.contains("flush_memwal") {
if path.contains("flush_rte") {
return accepted();
}
assert_eq!(path, "/v1/table/my_table/get_memwal_stats/");
assert_eq!(path, "/v1/table/my_table/get_rte_stats/");
ok_json(stats_body(&[], false))
});
table.checkpoint_memwal().await.unwrap();
table.checkpoint_rte().await.unwrap();
assert_eq!(compacts.load(std::sync::atomic::Ordering::SeqCst), 0);
}
@@ -10552,10 +10550,10 @@ mod tests {
let seen = compacts.clone();
let table = Table::new_with_handler("my_table", move |request| {
let path = request.url().path().to_string();
if path.contains("flush_memwal") {
if path.contains("flush_rte") {
return accepted();
}
if path.contains("compact_memwal") {
if path.contains("compact_rte") {
seen.fetch_add(1, std::sync::atomic::Ordering::SeqCst);
return accepted();
}
@@ -10565,7 +10563,7 @@ mod tests {
ok_json(stats_body(&left, false))
});
table.checkpoint_memwal().await.unwrap();
table.checkpoint_rte().await.unwrap();
assert_eq!(
compacts.load(std::sync::atomic::Ordering::SeqCst),
3,
@@ -10582,10 +10580,10 @@ mod tests {
let seen = compacts.clone();
let table = Table::new_with_handler("my_table", move |request| {
let path = request.url().path().to_string();
if path.contains("flush_memwal") {
if path.contains("flush_rte") {
return accepted();
}
if path.contains("compact_memwal") {
if path.contains("compact_rte") {
seen.fetch_add(1, std::sync::atomic::Ordering::SeqCst);
return accepted();
}
@@ -10599,7 +10597,7 @@ mod tests {
ok_json(body)
});
table.checkpoint_memwal().await.unwrap();
table.checkpoint_rte().await.unwrap();
assert_eq!(
compacts.load(std::sync::atomic::Ordering::SeqCst),
1,
@@ -10616,10 +10614,10 @@ mod tests {
let seen = compacts.clone();
let table = Table::new_with_handler("my_table", move |request| {
let path = request.url().path().to_string();
if path.contains("flush_memwal") {
if path.contains("flush_rte") {
return accepted();
}
if path.contains("compact_memwal") {
if path.contains("compact_rte") {
// First two triggers: every table shard already latched.
if seen.fetch_add(1, std::sync::atomic::Ordering::SeqCst) < 2 {
return http::Response::builder()
@@ -10641,7 +10639,7 @@ mod tests {
});
table
.checkpoint_memwal()
.checkpoint_rte()
.await
.expect("contention must not abort the checkpoint");
assert_eq!(
@@ -10661,7 +10659,7 @@ mod tests {
let seen = polls.clone();
let table = Table::new_with_handler("my_table", move |request| {
let path = request.url().path().to_string();
if path.contains("flush_memwal") || path.contains("compact_memwal") {
if path.contains("flush_rte") || path.contains("compact_rte") {
return accepted();
}
// The opening read lands; the next two polls are latched out.
@@ -10676,7 +10674,7 @@ mod tests {
});
table
.checkpoint_memwal()
.checkpoint_rte()
.await
.expect("a contended poll must be retried, not surfaced");
assert_eq!(
@@ -10696,7 +10694,7 @@ mod tests {
let seen = flushes.clone();
let table = Table::new_with_handler("my_table", move |request| {
let path = request.url().path().to_string();
if path.contains("flush_memwal") {
if path.contains("flush_rte") {
if seen.fetch_add(1, std::sync::atomic::Ordering::SeqCst) < 5 {
return http::Response::builder()
.status(429)
@@ -10705,14 +10703,14 @@ mod tests {
}
return accepted();
}
if path.contains("compact_memwal") {
if path.contains("compact_rte") {
return accepted();
}
ok_json(stats_body(&[], false))
});
table
.checkpoint_memwal()
.checkpoint_rte()
.await
.expect("contention must not be reported as a lost claim");
assert_eq!(
@@ -10737,7 +10735,7 @@ mod tests {
.unwrap()
});
let err = table.checkpoint_memwal().await.unwrap_err();
let err = table.checkpoint_rte().await.unwrap_err();
assert!(
matches!(&err, Error::Http { status_code: Some(s), .. } if s.as_u16() == 429),
"the fault that spent the budget must be the one reported: {err:?}"
@@ -10767,7 +10765,7 @@ mod tests {
.unwrap()
});
let err = table.checkpoint_memwal().await.unwrap_err();
let err = table.checkpoint_rte().await.unwrap_err();
let message = err.to_string();
assert!(
matches!(&err, Error::Http { status_code: Some(s), .. } if s.as_u16() == 503),
@@ -10795,7 +10793,7 @@ mod tests {
let seen = polls.clone();
let table = Table::new_with_handler("my_table", move |request| {
let path = request.url().path().to_string();
if path.contains("flush_memwal") || path.contains("compact_memwal") {
if path.contains("flush_rte") || path.contains("compact_rte") {
return accepted();
}
// Flat for far longer than any bound this loop ever had, with
@@ -10805,7 +10803,7 @@ mod tests {
});
table
.checkpoint_memwal()
.checkpoint_rte()
.await
.expect("a stall is the server being slow, not the client's call to make");
assert!(
@@ -10825,10 +10823,10 @@ mod tests {
let seen_compacts = compacts.clone();
let table = Table::new_with_handler("my_table", move |request| {
let path = request.url().path().to_string();
if path.contains("flush_memwal") {
if path.contains("flush_rte") {
return accepted();
}
if path.contains("compact_memwal") {
if path.contains("compact_rte") {
seen_compacts.fetch_add(1, std::sync::atomic::Ordering::SeqCst);
return accepted();
}
@@ -10842,7 +10840,7 @@ mod tests {
});
table
.checkpoint_memwal()
.checkpoint_rte()
.await
.expect("a running pass is progress, not a stall");
assert_eq!(
@@ -10856,9 +10854,9 @@ mod tests {
/// defaulting to a zero it did not measure. `include_sstable_rows`
/// rides in the body and is off unless asked for.
#[tokio::test]
async fn test_get_memwal_stats_round_trip() {
async fn test_get_rte_stats_round_trip() {
let table = Table::new_with_handler("my_table", |request| {
assert_eq!(request.url().path(), "/v1/table/my_table/get_memwal_stats/");
assert_eq!(request.url().path(), "/v1/table/my_table/get_rte_stats/");
let body = request.body().unwrap().as_bytes().unwrap();
let body: serde_json::Value = serde_json::from_slice(body).unwrap();
assert_eq!(
@@ -10866,7 +10864,7 @@ mod tests {
"the flag must reach the server, not be silently dropped"
);
let response = serde_json::json!({
"memwal_stats": {
"rte_stats": {
"table_shards": [{
"shard_id": "b0",
"status": "Active",
@@ -10891,7 +10889,7 @@ mod tests {
});
let stats = table
.get_memwal_stats(true)
.get_rte_stats(true)
.await
.unwrap()
.expect("a WAL-backed table reports Some");
@@ -10923,7 +10921,7 @@ mod tests {
.unwrap()
});
let err = table.checkpoint_memwal().await.unwrap_err();
let err = table.checkpoint_rte().await.unwrap_err();
assert!(
matches!(err, Error::TableNotFound { .. }),
"a missing table must say so: {err:?}"
@@ -10944,12 +10942,12 @@ mod tests {
let table = Table::new_with_handler("my_table", move |request| {
let path = request.url().path().to_string();
let n = seen.fetch_add(1, std::sync::atomic::Ordering::SeqCst);
if path.contains("flush_memwal") {
if path.contains("flush_rte") {
// First flush lands; the claim is then lost, and the
// re-issued flush succeeds.
return accepted();
}
if path.contains("compact_memwal") {
if path.contains("compact_rte") {
if n < 4 {
return http::Response::builder()
.status(421)
@@ -10962,20 +10960,20 @@ mod tests {
});
table
.checkpoint_memwal()
.checkpoint_rte()
.await
.expect("a lost claim must be recovered by re-flushing, not surfaced");
}
#[tokio::test]
async fn test_get_memwal_stats_absent_when_wal_off() {
async fn test_get_rte_stats_absent_when_wal_off() {
let table = Table::new_with_handler("my_table", |_request| {
http::Response::builder()
.status(200)
.body(serde_json::json!({ "memwal_stats": null }).to_string())
.body(serde_json::json!({ "rte_stats": null }).to_string())
.unwrap()
});
assert!(table.get_memwal_stats(false).await.unwrap().is_none());
assert!(table.get_rte_stats(false).await.unwrap().is_none());
}
#[tokio::test]
@@ -15006,9 +15004,10 @@ mod tests {
// WAL-PK-FUSION: delete everything from here to the end of `mod tests`.
/// Verbatim from sophon's `LSM_WITH_ROW_ID_UNSUPPORTED`; the fallback keys
/// off this sentence, so a test that paraphrased it would prove nothing.
const WAL_ROW_ID_REFUSAL_BODY: &str = r#"{"code":13,"error":"Bad request: the MemWAL LSM scanner does not support with_row_id (the LSM scanner exposes _rowaddr, not a stable _rowid); set use_memwal(false) to read the base table only (results will exclude un-compacted MemWAL data)"}"#;
/// Verbatim from sophon's `wal_unsupported(WITH_ROW_ID_UNSUPPORTED)`; the
/// fallback keys off this sentence, so a test that paraphrased it would prove
/// nothing.
const WAL_ROW_ID_REFUSAL_BODY: &str = r#"{"code":13,"error":"Bad request: WAL unsupported: the MemWAL fresh tier does not support with_row_id: it has no stable _rowid; set use_rte(false) to read the base table only (results will exclude un-compacted RTE data)"}"#;
use lance::dataset::ROW_ID;
@@ -15052,7 +15051,7 @@ mod tests {
RecordBatch::try_new(Arc::new(Schema::new(fields)), columns).unwrap()
}
/// A server that refuses `_rowid` the way a MemWAL table does, recording
/// A server that refuses `_rowid` the way an RTE table does, recording
/// every query body so a test can see which key each attempt asked for.
fn wal_refusing_table(bodies: Arc<std::sync::Mutex<Vec<serde_json::Value>>>) -> Table {
Table::new_with_handler("my_table", move |request| match request.url().path() {
@@ -15093,7 +15092,7 @@ mod tests {
.unwrap()
}
/// The optimistic path: ask for `_rowid`, and when a MemWAL table refuses,
/// The optimistic path: ask for `_rowid`, and when an RTE table refuses,
/// fall back to the primary key rather than failing.
#[tokio::test]
async fn test_hybrid_falls_back_to_the_primary_key_when_row_id_is_refused() {
@@ -15267,7 +15266,7 @@ mod tests {
// Translated, not passed through: the caller gets the reason and the
// way out, which a bare 400 does not carry.
assert!(matches!(err, Error::NotSupported { .. }), "{err:?}");
assert!(err.to_string().contains("use_memwal(false)"), "{err}");
assert!(err.to_string().contains("use_rte(false)"), "{err}");
assert!(
bodies
.lock()
rust/lancedb/src/remote/table/blobs.rs
+3 -3
View File
@@ -716,8 +716,8 @@ impl<S: HttpSend> RemoteTable<S> {
)),
select: Select::columns(&[column]),
with_row_id: true,
// Row ids name base-table rows, and the MemWAL scanner has no stable `_rowid`.
use_memwal: Some(false),
// Row ids name base-table rows, and the RTE scanner has no stable `_rowid`.
use_rte: Some(false),
..Default::default()
});
let body = self
@@ -1153,7 +1153,7 @@ mod tests {
.unwrap();
assert_eq!(body["columns"], serde_json::json!(["image"]));
assert_eq!(body["with_row_id"], serde_json::json!(true));
assert_eq!(body["use_memwal"], serde_json::json!(false));
assert_eq!(body["use_rte"], serde_json::json!(false));
let requested = filter_row_ids(body["filter"].as_str().unwrap());
captured.lock().unwrap().push(TakeMockRequest::Query {
row_ids: requested.clone(),
rust/lancedb/src/remote/table/insert.rs
+1 -1
View File
@@ -1497,7 +1497,7 @@ mod tests {
when_not_matched_by_source_delete: false,
when_not_matched_by_source_delete_filt: None,
use_index: true,
use_memwal: None,
use_rte: None,
};
let input: Arc<dyn ExecutionPlan> = Arc::new(ErroringExec::new());
rust/lancedb/src/table.rs
+157 -170
View File
@@ -80,12 +80,12 @@ pub mod datafusion;
pub(crate) mod dataset;
pub mod delete;
pub mod freshness;
pub mod memwal_stats;
pub mod merge;
pub mod optimize;
pub(crate) mod primary_key;
pub mod query;
pub mod refresh;
pub mod rte_stats;
pub mod schema_evolution;
pub mod update;
pub mod write_progress;
@@ -107,9 +107,9 @@ use futures::future::join_all;
pub use lance::dataset::refs::{BranchContents, Ref, TagContents, Tags as LanceTags};
pub use lance::dataset::scanner::DatasetRecordBatchStream;
pub use lance_index::optimize::OptimizeOptions;
pub use memwal_stats::{MemWalStats, MemtableStats, SsTableStats, TableShardStats};
pub use optimize::{CompactionOptions, OptimizeAction, OptimizeStats};
pub use refresh::RefreshColumnResult;
pub use rte_stats::{MemtableStats, RTEStats, SsTableStats, TableShardStats};
pub use schema_evolution::{
AddColumnsResult, AlterColumnsResult, DropColumnsResult, FieldMetadataUpdate,
UpdateFieldMetadataResult,
@@ -324,23 +324,23 @@ pub trait Tags: Send + Sync {
pub use self::merge::MergeResult;
/// Specification selecting Lance's MemWAL LSM-style write path for
/// `merge_insert`.
/// Specification enabling the realtime engine (RTE), which routes
/// `merge_insert` through Lance's MemWAL LSM-style write path.
///
/// Construct via [`MemWalWriteSpec::bucket`], [`MemWalWriteSpec::identity`], or
/// [`MemWalWriteSpec::unsharded`], then optionally chain
/// [`MemWalWriteSpec::with_maintained_indexes`] (indexes the MemWAL keeps up to
/// date) and [`MemWalWriteSpec::with_writer_config_defaults`] (default
/// Construct via [`RTEWriteSpec::bucket`], [`RTEWriteSpec::identity`], or
/// [`RTEWriteSpec::unsharded`], then optionally chain
/// [`RTEWriteSpec::with_maintained_indexes`] (indexes the RTE keeps up to
/// date) and [`RTEWriteSpec::with_writer_config_defaults`] (default
/// `ShardWriter` configuration recorded in the MemWAL index).
///
/// A fresh spec maintains every index on the table, including ones created
/// after it is installed.
///
/// Install a spec with [`Table::enable_memwal`] and remove it with
/// [`Table::disable_memwal`]. The actual `merge_insert` dispatch
/// onto the MemWAL writer is a follow-up.
/// Install a spec with [`Table::enable_rte`] and remove it with
/// [`Table::disable_rte`]. The actual `merge_insert` dispatch
/// onto the RTE writer is a follow-up.
#[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize)]
pub enum MemWalWriteSpec {
pub enum RTEWriteSpec {
/// Hash-bucket sharding by a scalar column.
///
/// `column` must be a non-nested column with a supported scalar type.
@@ -350,7 +350,7 @@ pub enum MemWalWriteSpec {
Bucket {
column: String,
num_buckets: u32,
/// Indexes the MemWAL maintains in-memory as rows are appended.
/// Indexes the RTE maintains in-memory as rows are appended.
///
/// `None` means every index the table has, including ones created
/// later. `Some([])` maintains nothing.
@@ -364,7 +364,7 @@ pub enum MemWalWriteSpec {
/// distinct value of `column` becomes its own shard.
Identity {
column: String,
/// Indexes the MemWAL maintains in-memory as rows are appended.
/// Indexes the RTE maintains in-memory as rows are appended.
///
/// `None` means every index the table has, including ones created
/// later. `Some([])` maintains nothing.
@@ -372,9 +372,9 @@ pub enum MemWalWriteSpec {
/// Default `ShardWriter` configuration recorded in the MemWAL index.
writer_config_defaults: HashMap<String, String>,
},
/// No sharding — every `merge_insert` call writes to a single MemWAL shard.
/// No sharding — every `merge_insert` call writes to a single RTE shard.
Unsharded {
/// Indexes the MemWAL maintains in-memory as rows are appended.
/// Indexes the RTE maintains in-memory as rows are appended.
///
/// `None` means every index the table has, including ones created
/// later. `Some([])` maintains nothing.
@@ -384,7 +384,7 @@ pub enum MemWalWriteSpec {
},
}
impl MemWalWriteSpec {
impl RTEWriteSpec {
/// Construct a hash-bucket sharding spec maintaining every index on the table.
pub fn bucket(column: impl Into<String>, num_buckets: u32) -> Self {
Self::Bucket {
@@ -400,7 +400,7 @@ impl MemWalWriteSpec {
///
/// `column` must be a deterministic function of the unenforced primary
/// key: every row with a given primary key must always produce the same
/// `column` value. MemWAL dedups upserts by primary key but tracks
/// `column` value. RTE dedups upserts by primary key but tracks
/// SSTables per shard, so if the same key is written with two
/// different `column` values its versions land in different shards and a
/// stale value can win. Typically `column` is the primary key itself, or
@@ -421,23 +421,23 @@ impl MemWalWriteSpec {
}
}
/// Set which indexes the MemWAL maintains.
/// Set which indexes the RTE maintains.
///
/// `None` (the default) is every index on the table, re-read as the table
/// changes, so an index created later is maintained too; one of a kind the
/// MemWAL cannot mirror is skipped rather than failing the table. A list is
/// RTE cannot mirror is skipped rather than failing the table. A list is
/// verbatim: each name must already exist and be maintainable, and an empty
/// list maintains nothing. The maintainable kinds are Lance's built-ins plus
/// any added with [`NativeTable::set_mem_index_registry`].
///
/// ```
/// # use lancedb::table::MemWalWriteSpec;
/// # use lancedb::table::RTEWriteSpec;
/// // Every index the table has, now and later:
/// MemWalWriteSpec::unsharded().with_maintained_indexes(None);
/// RTEWriteSpec::unsharded().with_maintained_indexes(None);
/// // Exactly these:
/// MemWalWriteSpec::unsharded().with_maintained_indexes(vec!["id_idx".to_string()]);
/// RTEWriteSpec::unsharded().with_maintained_indexes(vec!["id_idx".to_string()]);
/// // None at all:
/// MemWalWriteSpec::unsharded().with_maintained_indexes(Vec::new());
/// RTEWriteSpec::unsharded().with_maintained_indexes(Vec::new());
/// ```
pub fn with_maintained_indexes(mut self, indexes: impl Into<Option<Vec<String>>>) -> Self {
let indexes = indexes.into();
@@ -455,7 +455,7 @@ impl MemWalWriteSpec {
self
}
/// Replace the default `ShardWriter` configuration recorded in the MemWAL
/// Replace the default `ShardWriter` configuration recorded in the RTE
/// index, so every writer starts from the same defaults. Keys are
/// `ShardWriter` config field names (`Duration` knobs use a `_ms` suffix);
/// values are their string encodings.
@@ -486,7 +486,7 @@ impl MemWalWriteSpec {
self
}
/// Borrow the list of index names this spec asks MemWAL to maintain, or
/// Borrow the list of index names this spec asks RTE to maintain, or
/// `None` when it asks for every index on the table.
pub fn maintained_indexes(&self) -> Option<&[String]> {
match self {
@@ -674,37 +674,37 @@ pub trait BaseTable: std::fmt::Display + std::fmt::Debug + Send + Sync {
message: "set_unenforced_primary_key is not supported on this table type".into(),
})
}
/// Install an [`MemWalWriteSpec`] on this table.
/// Install an [`RTEWriteSpec`] on this table.
///
/// The spec selects Lance's MemWAL LSM-style write path for future
/// `merge_insert` calls.
///
/// The default implementation returns `NotSupported`. Implementations
/// that support the MemWAL write path must override this.
async fn enable_memwal(&self, _spec: MemWalWriteSpec) -> Result<()> {
/// that support the RTE write path must override this.
async fn enable_rte(&self, _spec: RTEWriteSpec) -> Result<()> {
Err(Error::NotSupported {
message: "enable_memwal is not supported on this table type".into(),
message: "enable_rte is not supported on this table type".into(),
})
}
/// Remove the [`MemWalWriteSpec`] from this table.
/// Remove the [`RTEWriteSpec`] from this table.
///
/// This is a no-op if no spec is currently set.
///
/// The default implementation returns `NotSupported`. Implementations
/// that support the MemWAL write path must override this.
async fn disable_memwal(&self) -> Result<()> {
/// that support the RTE write path must override this.
async fn disable_rte(&self) -> Result<()> {
Err(Error::NotSupported {
message: "disable_memwal is not supported on this table type".into(),
message: "disable_rte is not supported on this table type".into(),
})
}
/// Read the [`MemWalWriteSpec`] currently installed on this table, returning
/// `None` when the MemWAL write path is not enabled.
/// Read the [`RTEWriteSpec`] currently installed on this table, returning
/// `None` when the RTE write path is not enabled.
///
/// The default implementation returns `NotSupported`. Implementations that
/// support the MemWAL write path must override this.
async fn get_memwal_spec(&self) -> Result<Option<MemWalWriteSpec>> {
/// support the RTE write path must override this.
async fn get_rte_spec(&self) -> Result<Option<RTEWriteSpec>> {
Err(Error::NotSupported {
message: "get_memwal_spec is not supported on this table type".into(),
message: "get_rte_spec is not supported on this table type".into(),
})
}
/// Whether a hybrid query on this table has already been told it cannot
@@ -713,7 +713,7 @@ pub trait BaseTable: std::fmt::Display + std::fmt::Debug + Send + Sync {
/// WAL-PK-FUSION: delete this and `note_hybrid_pk_fusion`.
///
/// Learned, never probed: hybrid optimistically asks for `_rowid` and only
/// a MemWAL table refuses, so paying a round trip up front would tax every
/// an RTE table refuses, so paying a round trip up front would tax every
/// table to discover something almost none of them need. Synchronous and
/// free by construction — an implementation may only answer from what a
/// previous query already learned.
@@ -735,33 +735,33 @@ pub trait BaseTable: std::fmt::Display + std::fmt::Debug + Send + Sync {
/// Freeze every table shard's active memtable into an SSTable.
///
/// The default implementation returns `NotSupported`.
async fn flush_memwal(&self) -> Result<()> {
async fn flush_rte(&self) -> Result<()> {
Err(Error::NotSupported {
message: "flush_memwal is not supported on this table type".into(),
message: "flush_rte is not supported on this table type".into(),
})
}
/// Trigger a background SSTable compaction pass per table shard.
///
/// The default implementation returns `NotSupported`.
async fn compact_memwal(&self) -> Result<()> {
async fn compact_rte(&self) -> Result<()> {
Err(Error::NotSupported {
message: "compact_memwal is not supported on this table type".into(),
message: "compact_rte is not supported on this table type".into(),
})
}
/// Read live MemWAL state, or `None` when the MemWAL write path is not
/// Read live RTE state, or `None` when the RTE write path is not
/// enabled for this table.
///
/// The default implementation returns `NotSupported`.
async fn get_memwal_stats(&self, _include_sstable_rows: bool) -> Result<Option<MemWalStats>> {
async fn get_rte_stats(&self, _include_sstable_rows: bool) -> Result<Option<RTEStats>> {
Err(Error::NotSupported {
message: "get_memwal_stats is not supported on this table type".into(),
message: "get_rte_stats is not supported on this table type".into(),
})
}
/// Drain and close any cached MemWAL shard writers for this table.
/// Drain and close any cached RTE shard writers for this table.
///
/// The default implementation is a no-op; table types that maintain
/// MemWAL shard writers override it.
async fn close_memwal_writers(&self) -> Result<()> {
/// RTE shard writers override it.
async fn close_rte_writers(&self) -> Result<()> {
Ok(())
}
/// Names of the blob v2 columns in this table, in declaration order.
@@ -2008,53 +2008,53 @@ impl Table {
self.inner.set_unenforced_primary_key(&borrowed).await
}
/// Install an [`MemWalWriteSpec`] on this table, selecting Lance's MemWAL
/// Install an [`RTEWriteSpec`] on this table, selecting Lance's MemWAL
/// LSM-style write path for future `merge_insert` calls.
///
/// [`MemWalWriteSpec`] chooses one of three sharding strategies:
/// [`RTEWriteSpec`] chooses one of three sharding strategies:
///
/// - [`MemWalWriteSpec::bucket`] — hash-bucket writes by a scalar column.
/// - [`MemWalWriteSpec::identity`] — shard by the raw value of a scalar column.
/// - [`MemWalWriteSpec::unsharded`] — route every write to a single shard.
/// - [`RTEWriteSpec::bucket`] — hash-bucket writes by a scalar column.
/// - [`RTEWriteSpec::identity`] — shard by the raw value of a scalar column.
/// - [`RTEWriteSpec::unsharded`] — route every write to a single shard.
///
/// Every spec requires the table to declare an unenforced primary key
/// ([`Table::set_unenforced_primary_key`]); otherwise this fails.
///
/// A table carries one spec: this fails while one is installed, since the
/// SSTables already written were homed under it.
/// [`Table::disable_memwal`] removes one.
/// [`Table::disable_rte`] removes one.
///
/// # Example
///
/// ```
/// # use lancedb::table::{MemWalWriteSpec, Table};
/// # use lancedb::table::{RTEWriteSpec, Table};
/// # async fn example(table: &Table) -> Result<(), Box<dyn std::error::Error>> {
/// table
/// .enable_memwal(
/// MemWalWriteSpec::bucket("id", 16).with_maintained_indexes(vec!["id_idx".to_string()]),
/// .enable_rte(
/// RTEWriteSpec::bucket("id", 16).with_maintained_indexes(vec!["id_idx".to_string()]),
/// )
/// .await?;
/// # Ok(())
/// # }
/// ```
pub async fn enable_memwal(&self, spec: MemWalWriteSpec) -> Result<()> {
self.inner.enable_memwal(spec).await
pub async fn enable_rte(&self, spec: RTEWriteSpec) -> Result<()> {
self.inner.enable_rte(spec).await
}
/// Remove the [`MemWalWriteSpec`] from this table, reverting to the standard
/// Remove the [`RTEWriteSpec`] from this table, reverting to the standard
/// `merge_insert` write path.
///
/// Errors if no spec is currently set.
pub async fn disable_memwal(&self) -> Result<()> {
self.inner.disable_memwal().await
pub async fn disable_rte(&self) -> Result<()> {
self.inner.disable_rte().await
}
/// Read the [`MemWalWriteSpec`] currently installed on this table.
/// Read the [`RTEWriteSpec`] currently installed on this table.
///
/// `Ok(None)` means the MemWAL write path is not enabled at all — no spec has
/// been set, or one was removed with [`Table::disable_memwal`]. That
/// `Ok(None)` means the RTE write path is not enabled at all — no spec has
/// been set, or one was removed with [`Table::disable_rte`]. That
/// is a different answer from a spec whose
/// [`MemWalWriteSpec::maintained_indexes`] is `None`, which is an installed
/// [`RTEWriteSpec::maintained_indexes`] is `None`, which is an installed
/// spec selecting indexes automatically.
///
/// The spec read back is the one that was installed, selection included:
@@ -2067,21 +2067,21 @@ impl Table {
/// ```
/// # use lancedb::table::Table;
/// # async fn example(table: &Table) -> Result<(), Box<dyn std::error::Error>> {
/// if let Some(spec) = table.get_memwal_spec().await? {
/// println!("MemWAL write path enabled: {:?}", spec);
/// if let Some(spec) = table.get_rte_spec().await? {
/// println!("RTE write path enabled: {:?}", spec);
/// }
/// # Ok(())
/// # }
/// ```
pub async fn get_memwal_spec(&self) -> Result<Option<MemWalWriteSpec>> {
self.inner.get_memwal_spec().await
pub async fn get_rte_spec(&self) -> Result<Option<RTEWriteSpec>> {
self.inner.get_rte_spec().await
}
/// Converge this table's MemWAL write path into its base table.
/// Converge this table's RTE write path into its base table.
///
/// One `flush` to freeze every memtable into an SSTable, then compaction triggers
/// until every generation that existed at that moment has reached base.
/// The loop runs client-side, reading progress from `get_memwal_stats`, so
/// The loop runs client-side, reading progress from `get_rte_stats`, so
/// there is no held socket and nothing to reconcile if you drop this
/// future partway through.
///
@@ -2104,14 +2104,14 @@ impl Table {
/// ```no_run
/// # use lancedb::Table;
/// # async fn example(table: &Table) -> Result<(), Box<dyn std::error::Error>> {
/// let before = table.get_memwal_stats(false).await?;
/// table.checkpoint_memwal().await?;
/// let after = table.get_memwal_stats(false).await?;
/// let before = table.get_rte_stats(false).await?;
/// table.checkpoint_rte().await?;
/// let after = table.get_rte_stats(false).await?;
/// # Ok(())
/// # }
/// ```
pub async fn checkpoint_memwal(&self) -> Result<()> {
checkpoint::checkpoint_memwal(self).await
pub async fn checkpoint_rte(&self) -> Result<()> {
checkpoint::checkpoint_rte(self).await
}
/// Freeze every table shard's active memtable into an SSTable without touching the
@@ -2121,8 +2121,8 @@ impl Table {
/// a lower per-query cost. On a node that has not claimed this table it
/// claims it and replays the WAL log first — reporting "nothing to flush"
/// without replaying would lie about durable data.
pub async fn flush_memwal(&self) -> Result<()> {
self.inner.flush_memwal().await
pub async fn flush_rte(&self) -> Result<()> {
self.inner.flush_rte().await
}
/// Run one bounded SSTable compaction pass per table shard, reporting what
@@ -2130,11 +2130,11 @@ impl Table {
///
/// One pass, not convergence: that bounds each request's cost and gives a
/// caller driving its own cadence a progress signal per round trip.
pub async fn compact_memwal(&self) -> Result<()> {
self.inner.compact_memwal().await
pub async fn compact_rte(&self) -> Result<()> {
self.inner.compact_rte().await
}
/// Read live per-table-shard MemWAL state.
/// Read live per-table-shard RTE state.
///
/// Answers "how far behind is my fresh tier", "which table shard is hot", and
/// "why is my fresh-tier vector search brute-force". Mutates no table
@@ -2143,30 +2143,27 @@ impl Table {
///
/// `include_sstable_rows` reports a row count per SSTable. Off by
/// default: each count opens an uncached Lance dataset, and
/// `checkpoint_memwal` polls this needing only generation numbers.
/// `checkpoint_rte` polls this needing only generation numbers.
///
/// `Ok(None)` only when the MemWAL write path is not enabled, matching
/// [`Table::get_memwal_spec`]. Stats is fresh-tier only, so with the
/// `Ok(None)` only when the RTE write path is not enabled, matching
/// [`Table::get_rte_spec`]. Stats is fresh-tier only, so with the
/// WAL off there is no manifest to report and a struct of zeros would
/// read as measurements.
///
/// Do not build a checkpoint's termination on this: the completion
/// predicate lives in the `flush` and `compact` responses.
pub async fn get_memwal_stats(
&self,
include_sstable_rows: bool,
) -> Result<Option<MemWalStats>> {
self.inner.get_memwal_stats(include_sstable_rows).await
pub async fn get_rte_stats(&self, include_sstable_rows: bool) -> Result<Option<RTEStats>> {
self.inner.get_rte_stats(include_sstable_rows).await
}
/// Drain and close any cached MemWAL shard writers held for this table.
/// Drain and close any cached RTE shard writers held for this table.
///
/// When an [`MemWalWriteSpec`] is installed, `merge_insert` opens MemWAL shard
/// When an [`RTEWriteSpec`] is installed, `merge_insert` opens RTE shard
/// writers and caches them for reuse across calls. This closes them,
/// flushing pending data; writers reopen lazily on the next `merge_insert`.
/// It is a no-op when no writers are cached.
pub async fn close_memwal_writers(&self) -> Result<()> {
self.inner.close_memwal_writers().await
pub async fn close_rte_writers(&self) -> Result<()> {
self.inner.close_rte_writers().await
}
/// Retrieve the version of the table
@@ -2784,8 +2781,8 @@ impl NativeTable {
}
/// Maintain the memtable index kinds in `registry` when this table writes
/// through its [`MemWalWriteSpec`], and accept them when
/// [`MemWalWriteSpec::with_maintained_indexes`] names one.
/// through its [`RTEWriteSpec`], and accept them when
/// [`RTEWriteSpec::with_maintained_indexes`] names one.
///
/// The registry belongs to this handle, its clones and its branch handles,
/// and is not stored with the table: a handle opened separately starts from
@@ -2796,7 +2793,7 @@ impl NativeTable {
/// ```
/// # use std::sync::Arc;
/// # use lancedb::Table;
/// # use lancedb::table::{MemWalWriteSpec, MemIndexRegistry, mem_index::MemIndexPlugin};
/// # use lancedb::table::{RTEWriteSpec, MemIndexRegistry, mem_index::MemIndexPlugin};
/// # async fn example(
/// # table: &Table,
/// # plugin: Arc<dyn MemIndexPlugin>,
@@ -2805,8 +2802,8 @@ impl NativeTable {
/// if let Some(native) = table.as_native() {
/// native.set_mem_index_registry(registry);
/// }
/// let spec = MemWalWriteSpec::unsharded().with_maintained_indexes(vec!["region_bitmap".to_string()]);
/// table.enable_memwal(spec).await?;
/// let spec = RTEWriteSpec::unsharded().with_maintained_indexes(vec!["region_bitmap".to_string()]);
/// table.enable_rte(spec).await?;
/// # Ok(())
/// # }
/// ```
@@ -3699,20 +3696,20 @@ impl BaseTable for NativeTable {
primary_key::set_unenforced_primary_key(self, columns).await
}
async fn enable_memwal(&self, spec: MemWalWriteSpec) -> Result<()> {
merge::lsm::enable_memwal(self, spec).await
async fn enable_rte(&self, spec: RTEWriteSpec) -> Result<()> {
merge::lsm::enable_rte(self, spec).await
}
async fn disable_memwal(&self) -> Result<()> {
merge::lsm::disable_memwal(self).await
async fn disable_rte(&self) -> Result<()> {
merge::lsm::disable_rte(self).await
}
async fn get_memwal_spec(&self) -> Result<Option<MemWalWriteSpec>> {
merge::lsm::get_memwal_spec(self).await
async fn get_rte_spec(&self) -> Result<Option<RTEWriteSpec>> {
merge::lsm::get_rte_spec(self).await
}
async fn close_memwal_writers(&self) -> Result<()> {
merge::lsm::close_memwal_writers(self).await
async fn close_rte_writers(&self) -> Result<()> {
merge::lsm::close_rte_writers(self).await
}
async fn blob_columns(&self) -> Result<Vec<String>> {
@@ -5518,7 +5515,7 @@ mod tests {
}
#[tokio::test]
async fn test_enable_memwal() {
async fn test_enable_rte() {
use arrow_array::StringArray;
use lance::dataset::mem_wal::DatasetMemWalExt;
@@ -5551,15 +5548,15 @@ mod tests {
// Reject num_buckets out of range.
for bad in [0u32, 1025] {
let err = table
.enable_memwal(MemWalWriteSpec::bucket("id", bad))
.enable_rte(RTEWriteSpec::bucket("id", bad))
.await
.expect_err("should reject");
assert!(matches!(err, Error::InvalidInput { .. }), "got {:?}", err);
}
// Happy path: install spec; verify MemWAL details record it.
// Happy path: install spec; verify RTE details record it.
table
.enable_memwal(MemWalWriteSpec::bucket("id", 4))
.enable_rte(RTEWriteSpec::bucket("id", 4))
.await
.unwrap();
@@ -5585,7 +5582,7 @@ mod tests {
// Mutation rejected.
let err = table
.enable_memwal(MemWalWriteSpec::bucket("id", 8))
.enable_rte(RTEWriteSpec::bucket("id", 8))
.await
.expect_err("mutation should be rejected");
assert!(matches!(err, Error::InvalidInput { .. }), "got {:?}", err);
@@ -5594,7 +5591,7 @@ mod tests {
/// A spec is refused on a table without an unenforced primary key, for every
/// sharding mode, and a refused spec leaves nothing installed.
#[tokio::test]
async fn test_enable_memwal_requires_primary_key() {
async fn test_enable_rte_requires_primary_key() {
use lance::dataset::mem_wal::DatasetMemWalExt;
let tmp_dir = tempdir().unwrap();
@@ -5621,11 +5618,11 @@ mod tests {
};
for spec in [
MemWalWriteSpec::unsharded(),
MemWalWriteSpec::bucket("id", 4),
MemWalWriteSpec::identity("id"),
RTEWriteSpec::unsharded(),
RTEWriteSpec::bucket("id", 4),
RTEWriteSpec::identity("id"),
] {
let err = table.enable_memwal(spec).await.unwrap_err();
let err = table.enable_rte(spec).await.unwrap_err();
assert!(
matches!(&err, Error::InvalidInput { message } if message.contains("primary key")),
"got {err:?}"
@@ -5634,15 +5631,12 @@ mod tests {
}
table.set_unenforced_primary_key(["id"]).await.unwrap();
table
.enable_memwal(MemWalWriteSpec::unsharded())
.await
.unwrap();
table.enable_rte(RTEWriteSpec::unsharded()).await.unwrap();
assert!(installed(&table).await);
}
#[tokio::test]
async fn test_enable_memwal_unsharded() {
async fn test_enable_rte_unsharded() {
use lance::dataset::mem_wal::DatasetMemWalExt;
let tmp_dir = tempdir().unwrap();
@@ -5664,10 +5658,7 @@ mod tests {
let table = conn.create_table("t", reader).execute().await.unwrap();
table.set_unenforced_primary_key(["id"]).await.unwrap();
table
.enable_memwal(MemWalWriteSpec::unsharded())
.await
.unwrap();
table.enable_rte(RTEWriteSpec::unsharded()).await.unwrap();
let dataset = table.as_native().unwrap().dataset.get().await.unwrap();
let details = dataset
@@ -5683,7 +5674,7 @@ mod tests {
}
#[tokio::test]
async fn test_enable_memwal_identity() {
async fn test_enable_rte_identity() {
use lance::dataset::mem_wal::DatasetMemWalExt;
let tmp_dir = tempdir().unwrap();
@@ -5712,8 +5703,8 @@ mod tests {
table.set_unenforced_primary_key(["id"]).await.unwrap();
table
.enable_memwal(
MemWalWriteSpec::identity("region")
.enable_rte(
RTEWriteSpec::identity("region")
.with_writer_config_defaults([("durable_write", "false")]),
)
.await
@@ -5741,7 +5732,7 @@ mod tests {
}
#[tokio::test]
async fn test_disable_memwal() {
async fn test_disable_rte() {
use lance::dataset::mem_wal::DatasetMemWalExt;
let tmp_dir = tempdir().unwrap();
@@ -5764,11 +5755,11 @@ mod tests {
table.set_unenforced_primary_key(["id"]).await.unwrap();
// unset errors when no spec is set.
table.disable_memwal().await.unwrap_err();
table.disable_rte().await.unwrap_err();
// Install a spec, then unset it.
table
.enable_memwal(MemWalWriteSpec::bucket("id", 4))
.enable_rte(RTEWriteSpec::bucket("id", 4))
.await
.unwrap();
{
@@ -5776,16 +5767,16 @@ mod tests {
assert!(dataset.mem_wal_index_details().await.unwrap().is_some());
}
table.disable_memwal().await.unwrap();
table.disable_rte().await.unwrap();
{
let dataset = table.as_native().unwrap().dataset.get().await.unwrap();
assert!(dataset.mem_wal_index_details().await.unwrap().is_none());
}
// A second unset errors; a fresh spec can still be installed afterwards.
table.disable_memwal().await.unwrap_err();
table.disable_rte().await.unwrap_err();
table
.enable_memwal(MemWalWriteSpec::bucket("id", 8))
.enable_rte(RTEWriteSpec::bucket("id", 8))
.await
.unwrap();
{
@@ -5795,7 +5786,7 @@ mod tests {
}
#[tokio::test]
async fn test_get_memwal_spec() {
async fn test_get_rte_spec() {
let tmp_dir = tempdir().unwrap();
let uri = tmp_dir.path().to_str().unwrap();
@@ -5822,7 +5813,7 @@ mod tests {
table.set_unenforced_primary_key(["id"]).await.unwrap();
// No spec installed yet.
assert_eq!(table.get_memwal_spec().await.unwrap(), None);
assert_eq!(table.get_rte_spec().await.unwrap(), None);
// A real scalar index is needed to name it as a maintained index.
table
@@ -5834,35 +5825,35 @@ mod tests {
// Bucket spec round-trips exactly, including the routing column (recovered
// from its field id), maintained indexes, and writer config defaults.
let spec = MemWalWriteSpec::bucket("id", 4)
let spec = RTEWriteSpec::bucket("id", 4)
.with_maintained_indexes(vec![idx_name.clone()])
.with_writer_config_defaults([("durable_write", "false")]);
table.enable_memwal(spec.clone()).await.unwrap();
assert_eq!(table.get_memwal_spec().await.unwrap(), Some(spec));
table.enable_rte(spec.clone()).await.unwrap();
assert_eq!(table.get_rte_spec().await.unwrap(), Some(spec));
// After unset, no spec is reported.
table.disable_memwal().await.unwrap();
assert_eq!(table.get_memwal_spec().await.unwrap(), None);
table.disable_rte().await.unwrap();
assert_eq!(table.get_rte_spec().await.unwrap(), None);
// Identity sharding round-trips (column recovered from the schema).
// A spec left at its default round-trips as the default: what is stored
// is the intent to maintain everything, not the set it resolves to now.
let spec = MemWalWriteSpec::identity("region");
table.enable_memwal(spec.clone()).await.unwrap();
assert_eq!(table.get_memwal_spec().await.unwrap(), Some(spec));
table.disable_memwal().await.unwrap();
let spec = RTEWriteSpec::identity("region");
table.enable_rte(spec.clone()).await.unwrap();
assert_eq!(table.get_rte_spec().await.unwrap(), Some(spec));
table.disable_rte().await.unwrap();
// Unsharded round-trips (no routing column).
let spec = MemWalWriteSpec::unsharded();
table.enable_memwal(spec.clone()).await.unwrap();
assert_eq!(table.get_memwal_spec().await.unwrap(), Some(spec));
let spec = RTEWriteSpec::unsharded();
table.enable_rte(spec.clone()).await.unwrap();
assert_eq!(table.get_rte_spec().await.unwrap(), Some(spec));
}
/// The maintained set defaults to every index on the table. A named index
/// the memtable cannot build fails the install, because the caller asked
/// for it; an unnamed one is skipped, because it arrived by existing.
#[tokio::test]
async fn test_enable_memwal_infers_maintained_indexes() {
async fn test_enable_rte_infers_maintained_indexes() {
let tmp_dir = tempdir().unwrap();
let uri = tmp_dir.path().to_str().unwrap();
@@ -5903,9 +5894,8 @@ mod tests {
// Explicitly naming the bitmap index fails before anything commits.
let err = table
.enable_memwal(
MemWalWriteSpec::unsharded()
.with_maintained_indexes(vec!["tag_bitmap".to_string()]),
.enable_rte(
RTEWriteSpec::unsharded().with_maintained_indexes(vec!["tag_bitmap".to_string()]),
)
.await
.unwrap_err();
@@ -5913,18 +5903,15 @@ mod tests {
matches!(err, Error::InvalidInput { ref message } if message.contains("tag_bitmap")),
"expected the bitmap index to be rejected, got {err:?}"
);
assert_eq!(table.get_memwal_spec().await.unwrap(), None);
assert_eq!(table.get_rte_spec().await.unwrap(), None);
// The default names nothing, so the bitmap is skipped rather than
// refusing the table: it is maintained by existing, not by being asked
// for. The intent is what is stored, not the set it resolves to today.
table
.enable_memwal(MemWalWriteSpec::unsharded())
.await
.unwrap();
table.enable_rte(RTEWriteSpec::unsharded()).await.unwrap();
assert_eq!(
table
.get_memwal_spec()
.get_rte_spec()
.await
.unwrap()
.unwrap()
@@ -5935,16 +5922,16 @@ mod tests {
// Narrowing to a named subset: unset first, since an installed spec
// cannot be set over.
table.disable_memwal().await.unwrap();
table.disable_rte().await.unwrap();
table
.enable_memwal(
MemWalWriteSpec::unsharded().with_maintained_indexes(vec!["id_btree".to_string()]),
.enable_rte(
RTEWriteSpec::unsharded().with_maintained_indexes(vec!["id_btree".to_string()]),
)
.await
.unwrap();
assert_eq!(
table
.get_memwal_spec()
.get_rte_spec()
.await
.unwrap()
.unwrap()
@@ -5953,14 +5940,14 @@ mod tests {
);
// Opting out entirely is distinct from the default.
table.disable_memwal().await.unwrap();
table.disable_rte().await.unwrap();
table
.enable_memwal(MemWalWriteSpec::unsharded().with_maintained_indexes(Vec::new()))
.enable_rte(RTEWriteSpec::unsharded().with_maintained_indexes(Vec::new()))
.await
.unwrap();
assert_eq!(
table
.get_memwal_spec()
.get_rte_spec()
.await
.unwrap()
.unwrap()
@@ -5971,7 +5958,7 @@ mod tests {
// An installed spec is never set over, whatever the new one asks for,
// and the refused call leaves it alone.
let err = table
.enable_memwal(MemWalWriteSpec::bucket("id", 4))
.enable_rte(RTEWriteSpec::bucket("id", 4))
.await
.unwrap_err();
assert!(
@@ -5981,7 +5968,7 @@ mod tests {
);
assert_eq!(
table
.get_memwal_spec()
.get_rte_spec()
.await
.unwrap()
.unwrap()
rust/lancedb/src/table/checkpoint.rs
+12 -12
View File
@@ -1,12 +1,12 @@
// SPDX-License-Identifier: Apache-2.0
// SPDX-FileCopyrightText: Copyright The LanceDB Authors
//! Converging a table's MemWAL write path into its base table.
//! Converging a table's RTE write path into its base table.
//!
//! `checkpoint_memwal` freezes once, then triggers compaction and watches
//! `checkpoint_rte` freezes once, then triggers compaction and watches
//! generation numbers until the SSTables that existed at the start are gone.
//!
//! The loop runs in the client, not the server: `compact_memwal` dispatches a
//! The loop runs in the client, not the server: `compact_rte` dispatches a
//! pass and returns, so nothing holds a socket and a client can vanish
//! mid-operation with nothing to reconcile. Completion is read from
//! generation numbers in the shard manifest — durable state, unlike a count
@@ -68,7 +68,7 @@ fn is_lost_claim(e: &Error) -> bool {
status_of(e) == Some(421)
}
/// Interval between `get_memwal_stats` polls. One interval is roughly one
/// Interval between `get_rte_stats` polls. One interval is roughly one
/// compaction pass, the granularity at which the answer can change.
///
/// Fixed rather than configurable, matching `wait_for_index`. It costs
@@ -149,15 +149,15 @@ where
}
}
/// Drive [`Table::checkpoint_memwal`]: freeze once, fix the target watermark
/// Drive [`Table::checkpoint_rte`]: freeze once, fix the target watermark
/// from the resulting SSTables, then trigger and poll until they drain.
pub(crate) async fn checkpoint_memwal(table: &Table) -> Result<()> {
pub(crate) async fn checkpoint_rte(table: &Table) -> Result<()> {
for reissue in 0..=MAX_REISSUES {
// The freeze turns everything written before this call into an
// SSTable, so the watermark has to be read after it. Idempotent:
// freezing an empty memtable is a no-op, so a re-issue does not churn
// empty SSTables.
match issue(|| table.flush_memwal()).await? {
match issue(|| table.flush_rte()).await? {
Attempt::Ok(()) => {}
Attempt::ReissueFromFlush => {
backoff(reissue).await;
@@ -165,7 +165,7 @@ pub(crate) async fn checkpoint_memwal(table: &Table) -> Result<()> {
}
}
let stats = match issue(|| table.get_memwal_stats(false)).await? {
let stats = match issue(|| table.get_rte_stats(false)).await? {
Attempt::Ok(stats) => stats,
Attempt::ReissueFromFlush => {
backoff(reissue).await;
@@ -173,7 +173,7 @@ pub(crate) async fn checkpoint_memwal(table: &Table) -> Result<()> {
}
};
let Some(stats) = stats else {
// Not WAL-backed; `flush_memwal` would have errored first but for a race.
// Not WAL-backed; `flush_rte` would have errored first but for a race.
return Ok(());
};
let targets: HashMap<String, u64> = stats
@@ -194,7 +194,7 @@ pub(crate) async fn checkpoint_memwal(table: &Table) -> Result<()> {
}
}
Err(Error::Runtime {
message: "checkpoint_memwal: the owning node kept losing its claim; \
message: "checkpoint_rte: the owning node kept losing its claim; \
re-issued from flush the maximum number of times"
.into(),
})
@@ -213,7 +213,7 @@ async fn drain_to_targets(
targets: &HashMap<String, u64>,
) -> Result<CheckpointOutcome> {
loop {
let stats = match issue(|| table.get_memwal_stats(false)).await? {
let stats = match issue(|| table.get_rte_stats(false)).await? {
Attempt::Ok(stats) => stats,
Attempt::ReissueFromFlush => return Ok(CheckpointOutcome::ReissueFromFlush),
};
@@ -241,7 +241,7 @@ async fn drain_to_targets(
}
if !all_compacting {
match table.compact_memwal().await {
match table.compact_rte().await {
Ok(()) => {}
Err(e) if is_lost_claim(&e) => return Ok(CheckpointOutcome::ReissueFromFlush),
Err(e) if !is_retryable(&e) => return Err(e),
rust/lancedb/src/table/computed_columns.rs
+12 -18
View File
@@ -2197,7 +2197,7 @@ pub(crate) fn plan(schema: SchemaRef, columns: &[(String, String)]) -> Result<Ve
/// every declaration. For callers that stage declarations behind other work
/// and need those rejections before any of it lands.
///
/// Only the schema is consulted. Declaring also refuses a table with a MemWAL
/// Only the schema is consulted. Declaring also refuses a table with an RTE
/// write spec or retained SSTables; that is table state, checked at commit.
///
/// ```
@@ -3025,7 +3025,7 @@ mod tests {
/// The gate's reproducer: the reciprocal of the declare-under-spec check.
#[tokio::test]
async fn test_installing_an_lsm_spec_over_computed_columns_is_refused() {
use crate::table::MemWalWriteSpec;
use crate::table::RTEWriteSpec;
let tmp_dir = tempfile::tempdir().unwrap();
let conn = connect(tmp_dir.path().to_str().unwrap())
@@ -3053,14 +3053,14 @@ mod tests {
.unwrap();
let err = table
.enable_memwal(MemWalWriteSpec::unsharded())
.enable_rte(RTEWriteSpec::unsharded())
.await
.unwrap_err();
assert!(
matches!(&err, Error::NotSupported { message } if message.contains("computed")),
"{err:?}"
);
assert!(table.get_memwal_spec().await.unwrap().is_none());
assert!(table.get_rte_spec().await.unwrap().is_none());
}
/// The gate's reproducer: declaration metadata is admitted only through
@@ -3233,10 +3233,10 @@ mod tests {
);
}
/// Stale handles must not commit the computed/MemWAL state in either order.
/// Stale handles must not commit the computed/RTE state in either order.
#[tokio::test]
async fn test_stale_handles_cannot_mix_computed_and_lsm() {
use crate::table::MemWalWriteSpec;
use crate::table::RTEWriteSpec;
let tmp_dir = tempfile::tempdir().unwrap();
let uri = tmp_dir.path().to_str().unwrap();
@@ -3260,7 +3260,7 @@ mod tests {
.await
.unwrap();
let err = stale
.enable_memwal(MemWalWriteSpec::unsharded())
.enable_rte(RTEWriteSpec::unsharded())
.await
.unwrap_err();
assert!(matches!(err, Error::NotSupported { .. }), "install won");
@@ -3274,10 +3274,7 @@ mod tests {
let table = conn.create_table("mix2", batch).execute().await.unwrap();
table.set_unenforced_primary_key(["x"]).await.unwrap();
let stale = conn.open_table("mix2").execute().await.unwrap();
table
.enable_memwal(MemWalWriteSpec::unsharded())
.await
.unwrap();
table.enable_rte(RTEWriteSpec::unsharded()).await.unwrap();
let err = add_computed(&stale, &[("doubled".into(), "x * 2".into())])
.await
.unwrap_err();
@@ -3289,7 +3286,7 @@ mod tests {
/// flag is the durable marker; declaration refuses on it.
#[tokio::test]
async fn test_unset_with_retained_lsm_rows_cannot_admit_a_declaration() {
use crate::table::MemWalWriteSpec;
use crate::table::RTEWriteSpec;
use arrow_array::{Int64Array, RecordBatchIterator};
let tmp_dir = tempfile::tempdir().unwrap();
@@ -3315,21 +3312,18 @@ mod tests {
.await
.unwrap();
table.set_unenforced_primary_key(["id"]).await.unwrap();
table
.enable_memwal(MemWalWriteSpec::unsharded())
.await
.unwrap();
table.enable_rte(RTEWriteSpec::unsharded()).await.unwrap();
let mut merge = table.merge_insert(&["id"]);
merge
.when_matched_update_all(None)
.when_not_matched_insert_all()
.use_memwal(true);
.use_rte(true);
merge
.execute(Box::new(RecordBatchIterator::new(vec![Ok(batch)], schema)))
.await
.unwrap();
table.disable_memwal().await.unwrap();
table.disable_rte().await.unwrap();
let err = add_computed(&table, &[("doubled".into(), "value * 2".into())])
.await
rust/lancedb/src/table/datafusion.rs
+7 -7
View File
@@ -166,7 +166,7 @@ pub struct BaseTableAdapter {
table: Arc<dyn BaseTable>,
schema: Arc<ArrowSchema>,
fts_query: Option<FullTextSearchQuery>,
use_memwal: Option<bool>,
use_rte: Option<bool>,
}
impl BaseTableAdapter {
@@ -182,13 +182,13 @@ impl BaseTableAdapter {
table,
schema: Arc::new(schema),
fts_query: None,
use_memwal: None,
use_rte: None,
})
}
/// Set the MemWAL read routing of every scan; see [`QueryRequest::use_memwal`].
pub fn with_use_memwal(mut self, use_memwal: Option<bool>) -> Self {
self.use_memwal = use_memwal;
/// Set the RTE read routing of every scan; see [`QueryRequest::use_rte`].
pub fn with_use_rte(mut self, use_rte: Option<bool>) -> Self {
self.use_rte = use_rte;
self
}
@@ -204,7 +204,7 @@ impl BaseTableAdapter {
table: self.table.clone(),
schema,
fts_query: Some(fts_query),
use_memwal: self.use_memwal,
use_rte: self.use_rte,
}
}
}
@@ -232,7 +232,7 @@ impl TableProvider for BaseTableAdapter {
let mut query = QueryRequest {
full_text_search: self.fts_query.clone(),
disable_scoring_autoprojection: disable_scoring,
use_memwal: self.use_memwal,
use_rte: self.use_rte,
..Default::default()
};
rust/lancedb/src/table/dataset.rs
+5 -5
View File
@@ -19,7 +19,7 @@ use crate::{Error, error::Result, utils::background_cache::BackgroundCache};
pub struct DatasetConsistencyWrapper {
state: Arc<Mutex<DatasetState>>,
consistency: ConsistencyMode,
/// The single MemWAL `ShardWriter` for this dataset, co-located so it is
/// The single RTE `ShardWriter` for this dataset, co-located so it is
/// cached for the session and shares the dataset's lifecycle. A dataset
/// writes to one shard at a time. Shared by `Arc` across clones.
shard_writer: Arc<ShardWriterCache>,
@@ -33,7 +33,7 @@ struct DatasetState {
/// `None` = tracking latest.
pinned_version: Option<u64>,
/// Whether the pin is an internal query snapshot rather than user-visible
/// time travel. Query snapshots remain read-only but preserve MemWAL read
/// time travel. Query snapshots remain read-only but preserve RTE read
/// semantics.
query_snapshot: bool,
}
@@ -99,7 +99,7 @@ impl DatasetConsistencyWrapper {
}
/// Create an independent read-only wrapper pinned to the current dataset
/// while retaining this wrapper's live MemWAL read context.
/// while retaining this wrapper's live RTE read context.
pub async fn new_query_snapshot(&self) -> Result<Self> {
// Apply the configured consistency policy before taking the snapshot.
// The returned dataset is intentionally discarded: a checkout may race
@@ -109,7 +109,7 @@ impl DatasetConsistencyWrapper {
let (dataset, query_snapshot) = {
let state = self.state.lock()?;
// Preserve user time travel so the MemWAL safety guard still sees
// Preserve user time travel so the RTE safety guard still sees
// it. Latest and already-internal snapshots remain internal pins.
(
state.dataset.clone(),
@@ -128,7 +128,7 @@ impl DatasetConsistencyWrapper {
})
}
/// The MemWAL `ShardWriter` cache co-located with this dataset.
/// The RTE `ShardWriter` cache co-located with this dataset.
pub(crate) fn shard_writer(&self) -> &Arc<ShardWriterCache> {
&self.shard_writer
}
rust/lancedb/src/table/merge.rs
+76 -115
View File
@@ -47,7 +47,7 @@ pub struct MergeResult {
/// Total number of rows written.
///
/// On the standard `merge_insert` path this equals
/// `num_inserted_rows + num_updated_rows`. On the MemWAL write path the
/// `num_inserted_rows + num_updated_rows`. On the RTE write path the
/// insert/update breakdown is not known until compaction; in that mode
/// `num_inserted_rows`, `num_updated_rows`, `num_deleted_rows`, `version`
/// and `num_attempts` are all `0` and this field holds the total number of
@@ -76,7 +76,7 @@ pub struct MergeInsertBuilder {
pub(crate) when_not_matched_by_source_delete_filt: Option<MergeFilter>,
pub(crate) timeout: Option<Duration>,
pub(crate) use_index: bool,
pub(crate) use_memwal: Option<bool>,
pub(crate) use_rte: Option<bool>,
pub(crate) validate_single_shard: bool,
}
@@ -92,7 +92,7 @@ impl MergeInsertBuilder {
when_not_matched_by_source_delete_filt: None,
timeout: None,
use_index: true,
use_memwal: None,
use_rte: None,
validate_single_shard: true,
}
}
@@ -190,30 +190,30 @@ impl MergeInsertBuilder {
self
}
/// Control MemWAL routing for this `merge_insert`.
/// Control RTE routing for this `merge_insert`.
///
/// By default (unset), a `merge_insert` on a table with an
/// [`MemWalWriteSpec`](super::MemWalWriteSpec) installed is routed through Lance's
/// MemWAL shard writer; a table without one uses the standard path.
/// [`RTEWriteSpec`](super::RTEWriteSpec) installed is routed through Lance's
/// RTE shard writer; a table without one uses the standard path.
///
/// - `use_memwal(true)` forces MemWAL routing and errors if the table has no
/// MemWAL write spec.
/// - `use_memwal(false)` forces the standard write path even when a spec is set.
pub fn use_memwal(&mut self, enable: bool) -> &mut Self {
self.use_memwal = Some(enable);
/// - `use_rte(true)` forces RTE routing and errors if the table has no
/// RTE write spec.
/// - `use_rte(false)` forces the standard write path even when a spec is set.
pub fn use_rte(&mut self, enable: bool) -> &mut Self {
self.use_rte = Some(enable);
self
}
/// Controls how an LSM `merge_insert` checks that its input targets a
/// single shard.
///
/// When a table has a MemWAL write spec, every row in a `merge_insert` call
/// When a table has an RTE write spec, every row in a `merge_insert` call
/// must route to the same shard. When `true` (the default), every row is
/// inspected to verify this. When `false`, only the first row is inspected
/// and the shard it routes to is used for the whole input — a faster path
/// for callers that have already pre-sharded their input.
///
/// Has no effect on tables without a MemWAL write spec.
/// Has no effect on tables without an RTE write spec.
pub fn validate_single_shard(&mut self, validate_single_shard: bool) -> &mut Self {
self.validate_single_shard = validate_single_shard;
self
@@ -903,7 +903,7 @@ mod lsm_tests {
use crate::connect;
use crate::error::Error;
use crate::table::{MemWalWriteSpec, Table};
use crate::table::{RTEWriteSpec, Table};
/// A reader of `[id: Int64, value: Int64]` rows; `value` is `0..n`.
fn id_value_reader(ids: Vec<i64>) -> Box<dyn RecordBatchReader + Send> {
@@ -987,7 +987,7 @@ mod lsm_tests {
let table = id_value_table(&dir).await;
// num_buckets = 1: every row routes to the single bucket.
table
.enable_memwal(MemWalWriteSpec::bucket("id", 1))
.enable_rte(RTEWriteSpec::bucket("id", 1))
.await
.unwrap();
@@ -1001,7 +1001,7 @@ mod lsm_tests {
.await
.unwrap();
// LSM path: rows go to the MemWAL, the breakdown is unknown until
// LSM path: rows go to the RTE, the breakdown is unknown until
// compaction, so only `num_rows` is populated.
assert_eq!(result.num_rows, 3);
assert_eq!(result.version, 0);
@@ -1013,10 +1013,7 @@ mod lsm_tests {
async fn lsm_merge_insert_unsharded() {
let dir = tempdir().unwrap();
let table = id_value_table(&dir).await;
table
.enable_memwal(MemWalWriteSpec::unsharded())
.await
.unwrap();
table.enable_rte(RTEWriteSpec::unsharded()).await.unwrap();
let mut builder = table.merge_insert(&["id"]);
builder
@@ -1043,7 +1040,7 @@ mod lsm_tests {
.unwrap();
table.set_unenforced_primary_key(["id"]).await.unwrap();
table
.enable_memwal(MemWalWriteSpec::identity("region"))
.enable_rte(RTEWriteSpec::identity("region"))
.await
.unwrap();
@@ -1060,18 +1057,18 @@ mod lsm_tests {
}
#[tokio::test]
async fn lsm_merge_insert_use_memwal_false_falls_back() {
async fn lsm_merge_insert_use_rte_false_falls_back() {
let dir = tempdir().unwrap();
let table = id_value_table(&dir).await;
table
.enable_memwal(MemWalWriteSpec::bucket("id", 1))
.enable_rte(RTEWriteSpec::bucket("id", 1))
.await
.unwrap();
// use_memwal(false) opts out: the standard path runs and commits even though
// use_rte(false) opts out: the standard path runs and commits even though
// a spec is installed.
let mut builder = table.merge_insert(&["id"]);
builder.when_not_matched_insert_all().use_memwal(false);
builder.when_not_matched_insert_all().use_rte(false);
let result = builder
.execute(id_value_reader(vec![3, 4, 5]))
.await
@@ -1082,17 +1079,17 @@ mod lsm_tests {
}
#[tokio::test]
async fn lsm_merge_insert_use_memwal_true_without_spec_errors() {
async fn lsm_merge_insert_use_rte_true_without_spec_errors() {
let dir = tempdir().unwrap();
let table = id_value_table(&dir).await;
// use_memwal(true) demands MemWAL routing; without a write spec it errors
// use_rte(true) demands RTE routing; without a write spec it errors
// rather than silently falling back to the standard path.
let mut builder = table.merge_insert(&["id"]);
builder
.when_matched_update_all(None)
.when_not_matched_insert_all()
.use_memwal(true);
.use_rte(true);
let err = builder
.execute(id_value_reader(vec![3, 4, 5]))
.await
@@ -1105,7 +1102,7 @@ mod lsm_tests {
let dir = tempdir().unwrap();
let table = id_value_table(&dir).await;
table
.enable_memwal(MemWalWriteSpec::bucket("id", 1))
.enable_rte(RTEWriteSpec::bucket("id", 1))
.await
.unwrap();
@@ -1122,7 +1119,7 @@ mod lsm_tests {
let dir = tempdir().unwrap();
let table = id_value_table(&dir).await;
table
.enable_memwal(MemWalWriteSpec::bucket("id", 1))
.enable_rte(RTEWriteSpec::bucket("id", 1))
.await
.unwrap();
@@ -1138,7 +1135,7 @@ mod lsm_tests {
let dir = tempdir().unwrap();
let table = id_value_table(&dir).await;
table
.enable_memwal(MemWalWriteSpec::bucket("id", 1))
.enable_rte(RTEWriteSpec::bucket("id", 1))
.await
.unwrap();
@@ -1148,7 +1145,7 @@ mod lsm_tests {
.when_not_matched_insert_all();
builder.execute(id_value_reader(vec![7, 8])).await.unwrap();
table.close_memwal_writers().await.unwrap();
table.close_rte_writers().await.unwrap();
// The writer reopens lazily on the next merge_insert.
let mut builder = table.merge_insert(&[]);
@@ -1173,7 +1170,7 @@ mod lsm_tests {
.unwrap();
table.set_unenforced_primary_key(["id"]).await.unwrap();
table
.enable_memwal(MemWalWriteSpec::identity("region"))
.enable_rte(RTEWriteSpec::identity("region"))
.await
.unwrap();
@@ -1210,10 +1207,10 @@ mod lsm_tests {
#[tokio::test]
async fn lsm_merge_insert_no_spec_uses_standard_path() {
let dir = tempdir().unwrap();
// id_value_table sets a primary key but no MemWAL write spec.
// id_value_table sets a primary key but no RTE write spec.
let table = id_value_table(&dir).await;
// Without a spec, a default merge_insert (use_memwal unset) simply uses
// Without a spec, a default merge_insert (use_rte unset) simply uses
// the standard path and commits — no opt-out required, no error.
let mut builder = table.merge_insert(&["id"]);
builder
@@ -1241,7 +1238,7 @@ mod lsm_tests {
.unwrap();
table.set_unenforced_primary_key(["id"]).await.unwrap();
table
.enable_memwal(MemWalWriteSpec::identity("region"))
.enable_rte(RTEWriteSpec::identity("region"))
.await
.unwrap();
@@ -1267,7 +1264,7 @@ mod lsm_tests {
assert!(matches!(err, Error::InvalidInput { .. }), "got {err:?}");
// After closing the writer, a different shard can be written.
table.close_memwal_writers().await.unwrap();
table.close_rte_writers().await.unwrap();
let mut builder = table.merge_insert(&[]);
builder
.when_matched_update_all(None)
@@ -1279,7 +1276,7 @@ mod lsm_tests {
}
// ---------------------------------------------------------------------
// MemWAL read path
// RTE read path
// ---------------------------------------------------------------------
use crate::arrow::SendableRecordBatchStream;
@@ -1337,10 +1334,7 @@ mod lsm_tests {
async fn lsm_read_sees_active_memtable() {
let dir = tempdir().unwrap();
let table = id_value_table(&dir).await; // base: ids 1,2,3 (value 0,1,2)
table
.enable_memwal(MemWalWriteSpec::unsharded())
.await
.unwrap();
table.enable_rte(RTEWriteSpec::unsharded()).await.unwrap();
// Insert ids 4,5 into the active memtable (not committed to base).
lsm_upsert(&table, vec![4, 5]).await;
@@ -1353,8 +1347,8 @@ mod lsm_tests {
vec![1, 2, 3, 4, 5]
);
// use_memwal(false) bypasses the MemWAL and reads the base table only.
let base_only = table.query().use_memwal(false).execute().await.unwrap();
// use_rte(false) bypasses the RTE and reads the base table only.
let base_only = table.query().use_rte(false).execute().await.unwrap();
let rows = collect_id_value(base_only).await;
assert_eq!(
rows.iter().map(|(id, _)| *id).collect::<Vec<_>>(),
@@ -1363,23 +1357,20 @@ mod lsm_tests {
}
#[tokio::test]
async fn base_table_adapter_use_memwal_false_scans_the_base_only() {
async fn base_table_adapter_use_rte_false_scans_the_base_only() {
use crate::table::datafusion::BaseTableAdapter;
use datafusion::prelude::SessionContext;
let dir = tempdir().unwrap();
let table = id_value_table(&dir).await; // base: ids 1,2,3
table
.enable_memwal(MemWalWriteSpec::unsharded())
.await
.unwrap();
table.enable_rte(RTEWriteSpec::unsharded()).await.unwrap();
// Ids 4,5 land in the active memtable, not the base table.
lsm_upsert(&table, vec![4, 5]).await;
let adapter = BaseTableAdapter::try_new(table.base_table().clone())
.await
.unwrap()
.with_use_memwal(Some(false));
.with_use_rte(Some(false));
let ctx = SessionContext::new();
ctx.register_table("base_only", Arc::new(adapter)).unwrap();
let batches = ctx
@@ -1401,10 +1392,7 @@ mod lsm_tests {
async fn query_snapshot_preserves_lsm_read_semantics() {
let dir = tempdir().unwrap();
let table = id_value_table(&dir).await;
table
.enable_memwal(MemWalWriteSpec::unsharded())
.await
.unwrap();
table.enable_rte(RTEWriteSpec::unsharded()).await.unwrap();
lsm_upsert(&table, vec![4, 5]).await;
let snapshot = table.query_snapshot().await.unwrap();
@@ -1419,10 +1407,7 @@ mod lsm_tests {
async fn query_snapshot_preserves_time_travel_lsm_guard() {
let dir = tempdir().unwrap();
let table = id_value_table(&dir).await;
table
.enable_memwal(MemWalWriteSpec::unsharded())
.await
.unwrap();
table.enable_rte(RTEWriteSpec::unsharded()).await.unwrap();
lsm_upsert(&table, vec![4]).await;
let version = table.version().await.unwrap();
@@ -1439,10 +1424,7 @@ mod lsm_tests {
async fn lsm_read_dedup_newest_wins() {
let dir = tempdir().unwrap();
let table = id_value_table(&dir).await; // base: id 2 -> value 1
table
.enable_memwal(MemWalWriteSpec::unsharded())
.await
.unwrap();
table.enable_rte(RTEWriteSpec::unsharded()).await.unwrap();
// Upsert ids 2,3,4 with values 0,1,2. id 2 and 3 shadow the base rows.
lsm_upsert(&table, vec![2, 3, 4]).await;
@@ -1457,10 +1439,7 @@ mod lsm_tests {
async fn lsm_read_point_lookup_filter() {
let dir = tempdir().unwrap();
let table = id_value_table(&dir).await;
table
.enable_memwal(MemWalWriteSpec::unsharded())
.await
.unwrap();
table.enable_rte(RTEWriteSpec::unsharded()).await.unwrap();
lsm_upsert(&table, vec![2, 3, 4]).await; // id 2 -> value 0 (shadows base)
let lsm = table.query().only_if("id = 2").execute().await.unwrap();
@@ -1473,14 +1452,14 @@ mod lsm_tests {
let dir = tempdir().unwrap();
let table = id_value_table(&dir).await;
table
.enable_memwal(MemWalWriteSpec::bucket("id", 8))
.enable_rte(RTEWriteSpec::bucket("id", 8))
.await
.unwrap();
// Two single-row upserts that route to (likely) different shards; each
// closes the writer so the next opens a fresh shard.
lsm_upsert(&table, vec![10]).await;
table.close_memwal_writers().await.unwrap();
table.close_rte_writers().await.unwrap();
lsm_upsert(&table, vec![11]).await;
let lsm = table.query().execute().await.unwrap();
@@ -1494,15 +1473,12 @@ mod lsm_tests {
async fn lsm_read_after_close_sees_flushed() {
let dir = tempdir().unwrap();
let table = id_value_table(&dir).await;
table
.enable_memwal(MemWalWriteSpec::unsharded())
.await
.unwrap();
table.enable_rte(RTEWriteSpec::unsharded()).await.unwrap();
lsm_upsert(&table, vec![4, 5]).await;
// close flushes the active memtable to an on-disk generation and drops
// the cached writer; the read must still see those rows via the shard
// manifest snapshot.
table.close_memwal_writers().await.unwrap();
table.close_rte_writers().await.unwrap();
let lsm = table.query().execute().await.unwrap();
let ids: Vec<i64> = collect_id_value(lsm)
@@ -1516,11 +1492,11 @@ mod lsm_tests {
#[tokio::test]
async fn lsm_read_without_spec_reads_base() {
let dir = tempdir().unwrap();
let table = id_value_table(&dir).await; // no MemWAL write spec
let table = id_value_table(&dir).await; // no RTE write spec
// With no spec installed there is nothing to route: the default read and
// an explicit use_memwal(false) both read the base table without error.
for query in [table.query(), table.query().use_memwal(false)] {
// an explicit use_rte(false) both read the base table without error.
for query in [table.query(), table.query().use_rte(false)] {
let rows = collect_id_value(query.execute().await.unwrap()).await;
assert_eq!(
rows.iter().map(|(id, _)| *id).collect::<Vec<_>>(),
@@ -1530,16 +1506,13 @@ mod lsm_tests {
}
#[tokio::test]
async fn lsm_read_unsupported_shape_errors_without_use_memwal_false() {
async fn lsm_read_unsupported_shape_errors_without_use_rte_false() {
let dir = tempdir().unwrap();
let table = id_value_table(&dir).await;
table
.enable_memwal(MemWalWriteSpec::unsharded())
.await
.unwrap();
table.enable_rte(RTEWriteSpec::unsharded()).await.unwrap();
lsm_upsert(&table, vec![4]).await;
// `with_row_id` is a shape the LSM scanner cannot honor. On a MemWAL
// `with_row_id` is a shape the LSM scanner cannot honor. On an RTE
// table the default (auto-routed) read hard-errors rather than silently
// reading a stale base-only result that would exclude un-compacted row 4.
let err = table
@@ -1548,15 +1521,15 @@ mod lsm_tests {
.execute()
.await
.err()
.expect("unsupported shape on a MemWAL table must error");
.expect("unsupported shape on an RTE table must error");
assert!(matches!(err, Error::NotSupported { .. }), "got {err:?}");
// use_memwal(false) is the escape hatch: it reads the base table only.
// use_rte(false) is the escape hatch: it reads the base table only.
let rows = collect_id_value(
table
.query()
.with_row_id()
.use_memwal(false)
.use_rte(false)
.execute()
.await
.unwrap(),
@@ -1568,7 +1541,7 @@ mod lsm_tests {
);
}
/// A table with `id`/`text` and a primary key, ready for a MemWAL write spec.
/// A table with `id`/`text` and a primary key, ready for an RTE write spec.
async fn lsm_text_table(dir: &tempfile::TempDir) -> crate::Table {
let conn = connect(dir.path().to_str().unwrap())
.execute()
@@ -1583,7 +1556,7 @@ mod lsm_tests {
table
}
/// Upsert `rows` through the MemWAL write path.
/// Upsert `rows` through the RTE write path.
async fn upsert_text(table: &crate::Table, rows: Vec<(i64, &str)>) {
let mut builder = table.merge_insert(&[]);
builder
@@ -1626,10 +1599,7 @@ mod lsm_tests {
let dir = tempdir().unwrap();
let table = lsm_text_table(&dir).await;
// The default set, on a table with no FTS index yet.
table
.enable_memwal(MemWalWriteSpec::unsharded())
.await
.unwrap();
table.enable_rte(RTEWriteSpec::unsharded()).await.unwrap();
// Opens the writer, whose MemTable carries no FTS index.
upsert_text(&table, vec![(99, "zebra")]).await;
@@ -1661,7 +1631,7 @@ mod lsm_tests {
/// A spec is replaced by unsetting it and setting the new one.
///
/// `disable_memwal` drains the open writer, so the rows it held
/// `disable_rte` drains the open writer, so the rows it held
/// survive the replacement, and the writer the next write opens builds its
/// MemTables from the new set.
#[tokio::test]
@@ -1681,25 +1651,23 @@ mod lsm_tests {
// Installed maintaining nothing, then written to: the writer opens with
// no FTS index.
table
.enable_memwal(MemWalWriteSpec::unsharded().with_maintained_indexes(Vec::new()))
.enable_rte(RTEWriteSpec::unsharded().with_maintained_indexes(Vec::new()))
.await
.unwrap();
upsert_text(&table, vec![(99, "zebra")]).await;
// Setting over an installed spec is refused; unset, then set.
let err = table
.enable_memwal(
MemWalWriteSpec::unsharded().with_maintained_indexes(vec![fts_index.clone()]),
)
.enable_rte(RTEWriteSpec::unsharded().with_maintained_indexes(vec![fts_index.clone()]))
.await
.expect_err("an installed spec cannot be set over");
assert!(
err.to_string().contains("already set"),
"unexpected error: {err}"
);
table.disable_memwal().await.unwrap();
table.disable_rte().await.unwrap();
table
.enable_memwal(MemWalWriteSpec::unsharded().with_maintained_indexes(vec![fts_index]))
.enable_rte(RTEWriteSpec::unsharded().with_maintained_indexes(vec![fts_index]))
.await
.unwrap();
@@ -1739,10 +1707,7 @@ mod lsm_tests {
let dir = tempdir().unwrap();
let table = lsm_text_table(&dir).await;
// No index yet, and the default set: maintain whatever the table has.
table
.enable_memwal(MemWalWriteSpec::unsharded())
.await
.unwrap();
table.enable_rte(RTEWriteSpec::unsharded()).await.unwrap();
// Opens the writer, whose MemTable therefore carries no FTS index.
upsert_text(&table, vec![(99, "zebra")]).await;
@@ -1796,7 +1761,7 @@ mod lsm_tests {
.unwrap();
let fts_index = table.list_indices().await.unwrap()[0].name.clone();
table
.enable_memwal(MemWalWriteSpec::unsharded().with_maintained_indexes(vec![fts_index]))
.enable_rte(RTEWriteSpec::unsharded().with_maintained_indexes(vec![fts_index]))
.await
.unwrap();
@@ -1891,7 +1856,7 @@ mod lsm_tests {
}
let vec_index = table.list_indices().await.unwrap()[0].name.clone();
table
.enable_memwal(MemWalWriteSpec::unsharded().with_maintained_indexes(vec![vec_index]))
.enable_rte(RTEWriteSpec::unsharded().with_maintained_indexes(vec![vec_index]))
.await
.unwrap();
@@ -2163,11 +2128,10 @@ mod lsm_tests {
.execute()
.await
.unwrap();
let spec = || {
MemWalWriteSpec::unsharded().with_maintained_indexes(vec!["region_bitmap".to_string()])
};
let spec =
|| RTEWriteSpec::unsharded().with_maintained_indexes(vec!["region_bitmap".to_string()]);
let refused = table.enable_memwal(spec()).await.unwrap_err();
let refused = table.enable_rte(spec()).await.unwrap_err();
assert!(
refused.to_string().contains("region_bitmap")
&& refused
@@ -2180,7 +2144,7 @@ mod lsm_tests {
.with_plugin(Arc::new(BitmapAsBTree))
.unwrap();
table.as_native().unwrap().set_mem_index_registry(registry);
table.enable_memwal(spec()).await.unwrap();
table.enable_rte(spec()).await.unwrap();
// The registry is not stored with the table.
let other = conn.open_table("t").execute().await.unwrap();
@@ -2238,7 +2202,7 @@ mod lsm_tests {
active.index_store.index_names()
);
table.close_memwal_writers().await.unwrap();
table.close_rte_writers().await.unwrap();
assert_eq!(fresh().await, vec![10, 11]);
assert_sstables_answer_through_region_bitmap(&table, vec![10, 11]).await;
}
@@ -2345,10 +2309,7 @@ mod lsm_tests {
.is_some(),
"a clone shares the handle's registry"
);
table
.enable_memwal(MemWalWriteSpec::unsharded())
.await
.unwrap();
table.enable_rte(RTEWriteSpec::unsharded()).await.unwrap();
let upsert = |rows: Vec<(i64, &'static str)>| {
let table = table.clone();
@@ -2389,7 +2350,7 @@ mod lsm_tests {
active.index_store.index_names()
);
table.close_memwal_writers().await.unwrap();
table.close_rte_writers().await.unwrap();
// Only the SSTable written after the index exists carries it.
let sstables = published_sstables(&table).await;
let mut carrying = 0;
rust/lancedb/src/table/merge/lsm.rs
+68 -68
View File
@@ -1,10 +1,10 @@
// SPDX-License-Identifier: Apache-2.0
// SPDX-FileCopyrightText: Copyright The LanceDB Authors
//! MemWAL write path for `merge_insert`.
//! RTE write path for `merge_insert`.
//!
//! [`enable_memwal`] installs an [`MemWalWriteSpec`] on a table by creating
//! Lance's MemWAL index; [`disable_memwal`] removes it. Once a spec is
//! [`enable_rte`] installs an [`RTEWriteSpec`] on a table by creating
//! Lance's MemWAL index; [`disable_rte`] removes it. Once a spec is
//! installed, `merge_insert` upsert calls are dispatched through Lance's MemWAL
//! `ShardWriter` (LSM-style append) instead of the standard merge path — see
//! [`lsm_dispatch_decision`] and [`execute_lsm_merge_insert`].
@@ -14,7 +14,7 @@
//! unsharded). [`MergeInsertBuilder::validate_single_shard`] controls whether
//! every row is checked or only the first. A dataset writes to one shard at a
//! time; its writer is cached in the [`ShardWriterCache`] held alongside the
//! dataset, and [`close_memwal_writers`] closes it.
//! dataset, and [`close_rte_writers`] closes it.
use std::collections::HashMap;
use std::sync::Arc;
@@ -42,9 +42,9 @@ use uuid::Uuid;
use crate::error::{Error, Result};
use crate::index::IndexConfig;
use crate::table::merge::{MergeInsertBuilder, MergeResult};
use crate::table::{BaseTable, MemWalWriteSpec, NativeTable};
use crate::table::{BaseTable, NativeTable, RTEWriteSpec};
/// Spec id of the sole sharding spec installed by [`enable_memwal`].
/// Spec id of the sole sharding spec installed by [`enable_rte`].
/// Must match Lance's `InitializeMemWalBuilder` (`SHARDING_SPEC_ID`).
const SHARDING_SPEC_ID: u32 = 1;
@@ -63,23 +63,23 @@ const NUM_BUCKETS_PARAM: &str = "num_buckets";
const SHARD_NAMESPACE: Uuid = Uuid::from_u128(0x4c53_4d57_5249_5445_5f53_4841_5244_3031);
// =============================================================================
// enable_memwal
// enable_rte
// =============================================================================
/// Install an [`MemWalWriteSpec`] on the table.
/// Install an [`RTEWriteSpec`] on the table.
///
/// The bucket / identity / unsharded sharding spec is constructed and validated
/// by Lance's
/// [`InitializeMemWalBuilder`](lance::dataset::mem_wal::InitializeMemWalBuilder).
#[allow(clippy::redundant_pub_crate)]
pub(crate) async fn enable_memwal(table: &NativeTable, spec: MemWalWriteSpec) -> Result<()> {
pub(crate) async fn enable_rte(table: &NativeTable, spec: RTEWriteSpec) -> Result<()> {
table.dataset.ensure_mutable()?;
{
let dataset = table.dataset.get().await?;
if dataset.mem_wal_index_details().await?.is_some() {
return Err(Error::InvalidInput {
message: "enable_memwal: a MemWAL write spec is already set on this \
message: "enable_rte: an RTE write spec is already set on this \
table and cannot be changed"
.into(),
});
@@ -117,25 +117,25 @@ pub(crate) async fn enable_memwal(table: &NativeTable, spec: MemWalWriteSpec) ->
let schema = arrow_schema::Schema::from(dataset.schema());
if !crate::table::computed_columns::computed_columns(&schema).is_empty() {
return Err(Error::NotSupported {
message: "a MemWAL write spec cannot be installed on a table with computed \
message: "an RTE write spec cannot be installed on a table with computed \
columns: rows in un-compacted tiers are invisible to refresh"
.into(),
});
}
if crate::materialized_view::read_definition(&dataset.schema().metadata)?.is_some() {
return Err(Error::NotSupported {
message: "a MemWAL write spec cannot be installed on a materialized view: \
message: "an RTE write spec cannot be installed on a materialized view: \
rows in un-compacted tiers are invisible to refresh"
.into(),
});
}
// Every MemWAL write and read resolves rows by the unenforced primary key,
// Every RTE write and read resolves rows by the unenforced primary key,
// so a spec on a table without one leaves the table unwritable through the
// MemWAL path. A key cannot be removed once set, and an overwrite or restore
// RTE path. A key cannot be removed once set, and an overwrite or restore
// of the table conflicts with the commit below.
if dataset.schema().unenforced_primary_key().is_empty() {
return Err(Error::InvalidInput {
message: "a MemWAL write spec requires the table to declare an unenforced \
message: "an RTE write spec requires the table to declare an unenforced \
primary key; set one with set_unenforced_primary_key first"
.into(),
});
@@ -144,7 +144,7 @@ pub(crate) async fn enable_memwal(table: &NativeTable, spec: MemWalWriteSpec) ->
.initialize_mem_wal()
.mem_index_registry(table.mem_index_registry());
let writer_config_defaults = match spec {
MemWalWriteSpec::Bucket {
RTEWriteSpec::Bucket {
column,
num_buckets,
writer_config_defaults,
@@ -153,7 +153,7 @@ pub(crate) async fn enable_memwal(table: &NativeTable, spec: MemWalWriteSpec) ->
builder = builder.bucket_sharding(column, num_buckets);
writer_config_defaults
}
MemWalWriteSpec::Identity {
RTEWriteSpec::Identity {
column,
writer_config_defaults,
..
@@ -161,7 +161,7 @@ pub(crate) async fn enable_memwal(table: &NativeTable, spec: MemWalWriteSpec) ->
builder = builder.identity_sharding(column);
writer_config_defaults
}
MemWalWriteSpec::Unsharded {
RTEWriteSpec::Unsharded {
writer_config_defaults,
..
} => {
@@ -191,22 +191,22 @@ fn index_name_list(indices: &[IndexConfig]) -> String {
}
// =============================================================================
// disable_memwal
// disable_rte
// =============================================================================
/// Remove the [`MemWalWriteSpec`] from the table by dropping the MemWAL index.
/// Remove the [`RTEWriteSpec`] from the table by dropping the MemWAL index.
///
/// Any cached shard writers are drained and closed first. Errors if no spec is
/// currently set.
#[allow(clippy::redundant_pub_crate)]
pub(crate) async fn disable_memwal(table: &NativeTable) -> Result<()> {
pub(crate) async fn disable_rte(table: &NativeTable) -> Result<()> {
table.dataset.ensure_mutable()?;
{
let dataset = table.dataset.get().await?;
if dataset.mem_wal_index_details().await?.is_none() {
return Err(Error::InvalidInput {
message: "disable_memwal: no MemWAL write spec is set on this table".into(),
message: "disable_rte: no RTE write spec is set on this table".into(),
});
}
}
@@ -222,46 +222,46 @@ pub(crate) async fn disable_memwal(table: &NativeTable) -> Result<()> {
}
// =============================================================================
// get_memwal_spec
// get_rte_spec
// =============================================================================
/// Read the [`MemWalWriteSpec`] currently installed on the table, if any.
/// Read the [`RTEWriteSpec`] currently installed on the table, if any.
///
/// Reconstructs the spec from the MemWAL index's stored details — the same
/// metadata [`enable_memwal`] writes — read directly via
/// metadata [`enable_rte`] writes — read directly via
/// [`mem_wal_index_details`](DatasetMemWalExt::mem_wal_index_details) (the raw
/// manifest record, not the `describe_indices` enrichment path, so it is
/// unaffected by system-index filtering there). Returns `Ok(None)` when no
/// MemWAL index is installed.
#[allow(clippy::redundant_pub_crate)]
pub(crate) async fn get_memwal_spec(table: &NativeTable) -> Result<Option<MemWalWriteSpec>> {
pub(crate) async fn get_rte_spec(table: &NativeTable) -> Result<Option<RTEWriteSpec>> {
let dataset = table.dataset.get().await?;
let Some(details) = dataset.mem_wal_index_details().await? else {
return Ok(None);
};
let spec = memwal_spec_from_details(&details, dataset.schema())?;
let spec = rte_spec_from_details(&details, dataset.schema())?;
Ok(Some(spec))
}
/// Reconstruct the public [`MemWalWriteSpec`] from stored MemWAL index details.
/// Reconstruct the public [`RTEWriteSpec`] from stored MemWAL index details.
///
/// The sharding transform and its parameters recover the mode; the routing
/// column is resolved from the sharding field's source id back to its schema
/// name (only bucket / identity carry a column). `maintained_indexes` and
/// `writer_config_defaults` are copied verbatim, so the result round-trips
/// what `enable_memwal` installed.
fn memwal_spec_from_details(
/// what `enable_rte` installed.
fn rte_spec_from_details(
details: &MemWalIndexDetails,
schema: &LanceSchema,
) -> Result<MemWalWriteSpec> {
) -> Result<RTEWriteSpec> {
let spec = details
.sharding_specs
.first()
.ok_or_else(|| Error::Runtime {
message: "get_memwal_spec: MemWAL index has no sharding spec".to_string(),
message: "get_rte_spec: MemWAL index has no sharding spec".to_string(),
})?;
let field = spec.fields.first().ok_or_else(|| Error::Runtime {
message: "get_memwal_spec: MemWAL index has an empty sharding spec".to_string(),
message: "get_rte_spec: MemWAL index has an empty sharding spec".to_string(),
})?;
let base = match field.transform.as_deref() {
@@ -272,16 +272,16 @@ fn memwal_spec_from_details(
.and_then(|s| s.parse::<u32>().ok())
.filter(|n| *n > 0)
.ok_or_else(|| Error::Runtime {
message: "get_memwal_spec: MemWAL bucket spec has a missing or invalid num_buckets parameter".to_string(),
message: "get_rte_spec: RTE bucket spec has a missing or invalid num_buckets parameter".to_string(),
})?;
MemWalWriteSpec::bucket(sharding_column(field, schema)?, num_buckets)
RTEWriteSpec::bucket(sharding_column(field, schema)?, num_buckets)
}
Some(IDENTITY_TRANSFORM) => MemWalWriteSpec::identity(sharding_column(field, schema)?),
Some(UNSHARDED_TRANSFORM) => MemWalWriteSpec::unsharded(),
Some(IDENTITY_TRANSFORM) => RTEWriteSpec::identity(sharding_column(field, schema)?),
Some(UNSHARDED_TRANSFORM) => RTEWriteSpec::unsharded(),
other => {
return Err(Error::Runtime {
message: format!(
"get_memwal_spec: MemWAL index has an unsupported sharding transform {:?}",
"get_rte_spec: MemWAL index has an unsupported sharding transform {:?}",
other
),
});
@@ -297,30 +297,30 @@ fn memwal_spec_from_details(
/// Resolve the single routing column name from a sharding field's source id.
///
/// `enable_memwal` records the shard column by its Lance field id, so the
/// `enable_rte` records the shard column by its Lance field id, so the
/// name is recovered by looking that id up in the current schema.
fn sharding_column(field: &ShardingField, schema: &LanceSchema) -> Result<String> {
let source_id = *field.source_ids.first().ok_or_else(|| Error::Runtime {
message: "get_memwal_spec: sharding field has no source column".to_string(),
message: "get_rte_spec: sharding field has no source column".to_string(),
})?;
schema
.field_by_id(source_id)
.map(|f| f.name.clone())
.ok_or_else(|| Error::Runtime {
message: format!(
"get_memwal_spec: sharding source field id {} not found in schema",
"get_rte_spec: sharding source field id {} not found in schema",
source_id
),
})
}
// =============================================================================
// close_memwal_writers
// close_rte_writers
// =============================================================================
/// Drain and close every cached MemWAL shard writer for the table.
/// Drain and close every cached RTE shard writer for the table.
#[allow(clippy::redundant_pub_crate)]
pub(crate) async fn close_memwal_writers(table: &NativeTable) -> Result<()> {
pub(crate) async fn close_rte_writers(table: &NativeTable) -> Result<()> {
table.dataset.shard_writer().drain_and_close().await
}
@@ -328,13 +328,13 @@ pub(crate) async fn close_memwal_writers(table: &NativeTable) -> Result<()> {
// ShardWriter cache
// =============================================================================
/// Per-table cache holding the single open MemWAL `ShardWriter`.
/// Per-table cache holding the single open RTE `ShardWriter`.
///
/// Held by [`DatasetConsistencyWrapper`](crate::table::dataset::DatasetConsistencyWrapper)
/// so the writer lives where the dataset lives — cached for the session and
/// reused across `merge_insert` calls. A dataset writes to one shard at a
/// time; routing a `merge_insert` to a different shard requires closing the
/// current writer first via [`close_memwal_writers`]. `ShardWriter::put` takes
/// current writer first via [`close_rte_writers`]. `ShardWriter::put` takes
/// `&self`, so concurrent puts on the cached writer are safe; `close` consumes
/// the writer, so the entry wraps it in `RwLock<Option<ShardWriter>>`.
#[derive(Default)]
@@ -384,7 +384,7 @@ impl ShardWriterEntry {
/// The cached writer's latest in-memory manifest (current generation +
/// SSTables). `Ok(None)` if the writer was already closed.
/// Used by the MemWAL read path to snapshot this shard authoritatively
/// Used by the RTE read path to snapshot this shard authoritatively
/// without re-reading the on-disk manifest.
async fn manifest(&self) -> Result<Option<lance_index::mem_wal::ShardManifest>> {
let guard = self.inner.read().await;
@@ -444,7 +444,7 @@ impl ShardWriterCache {
.await
.map_err(|e| Error::Runtime {
message: format!(
"merge_insert: failed to open MemWAL shard writer for shard {}: {}",
"merge_insert: failed to open RTE shard writer for shard {}: {}",
shard_id, e
),
})?;
@@ -502,7 +502,7 @@ impl ShardWriterCache {
}
}
/// Snapshot the cached writer's shard for the MemWAL read path: its shard id,
/// Snapshot the cached writer's shard for the RTE read path: its shard id,
/// authoritative in-memory manifest, and active + frozen memtable refs.
/// Returns `None` when no writer is currently cached (e.g. nothing has been
/// written this session, or the writer was closed).
@@ -550,7 +550,7 @@ fn check_shard_match(cached: Uuid, wanted: Uuid) -> Result<()> {
}
Err(Error::InvalidInput {
message: format!(
"merge_insert: a shard writer is already open for shard {} but this input routes to shard {}; call close_memwal_writers before writing to a different shard",
"merge_insert: a shard writer is already open for shard {} but this input routes to shard {}; call close_rte_writers before writing to a different shard",
cached, wanted
),
})
@@ -571,7 +571,7 @@ enum LsmMode {
Unsharded,
}
/// Resolved plan for routing a `merge_insert` through the MemWAL write path.
/// Resolved plan for routing a `merge_insert` through the RTE write path.
#[derive(Debug)]
#[allow(clippy::redundant_pub_crate)]
pub(crate) struct LsmPlan {
@@ -582,13 +582,13 @@ pub(crate) struct LsmPlan {
/// Outcome of [`lsm_dispatch_decision`].
#[allow(clippy::redundant_pub_crate)]
pub(crate) enum LsmDispatch {
/// No MemWAL write spec applies; use the standard `merge_insert` path.
/// No RTE write spec applies; use the standard `merge_insert` path.
Standard,
/// Route the `merge_insert` through the MemWAL shard writer.
/// Route the `merge_insert` through the RTE shard writer.
Lsm(LsmPlan),
}
/// Decide whether a `merge_insert` should be routed through the MemWAL write
/// Decide whether a `merge_insert` should be routed through the RTE write
/// path, validating the builder against the installed spec.
#[allow(clippy::redundant_pub_crate)]
pub(crate) async fn lsm_dispatch_decision(
@@ -596,18 +596,18 @@ pub(crate) async fn lsm_dispatch_decision(
params: &MergeInsertBuilder,
) -> Result<LsmDispatch> {
// Explicit opt-out: use the standard path regardless of any installed spec.
if params.use_memwal == Some(false) {
if params.use_rte == Some(false) {
return Ok(LsmDispatch::Standard);
}
let dataset = table.dataset.get().await?;
let Some(details) = dataset.mem_wal_index_details().await? else {
// No write spec installed. `use_memwal(true)` demanded MemWAL routing, so
// No write spec installed. `use_rte(true)` demanded RTE routing, so
// that is an error; otherwise fall back to the standard path.
if params.use_memwal == Some(true) {
if params.use_rte == Some(true) {
return Err(Error::InvalidInput {
message: "use_memwal(true) was set but the table has no MemWAL write spec; \
install one with enable_memwal or leave use_memwal unset"
message: "use_rte(true) was set but the table has no RTE write spec; \
install one with enable_rte or leave use_rte unset"
.to_string(),
});
}
@@ -629,7 +629,7 @@ pub(crate) async fn lsm_dispatch_decision(
if !params.on.is_empty() && params.on != pk_cols {
return Err(Error::InvalidInput {
message: format!(
"merge_insert: `on` columns {:?} must match the table's unenforced primary key {:?} when a MemWAL write spec is set; pass an empty `on` to default to the primary key",
"merge_insert: `on` columns {:?} must match the table's unenforced primary key {:?} when an RTE write spec is set; pass an empty `on` to default to the primary key",
params.on, pk_cols
),
});
@@ -637,7 +637,7 @@ pub(crate) async fn lsm_dispatch_decision(
if !is_upsert_only(params) {
return Err(Error::InvalidInput {
message: "merge_insert: when a MemWAL write spec is set, only the upsert form (when_matched_update_all without a filter + when_not_matched_insert_all, no by-source delete) is supported; call use_memwal(false) to use the standard merge_insert path".to_string(),
message: "merge_insert: when an RTE write spec is set, only the upsert form (when_matched_update_all without a filter + when_not_matched_insert_all, no by-source delete) is supported; call use_rte(false) to use the standard merge_insert path".to_string(),
});
}
@@ -678,7 +678,7 @@ fn resolve_lsm_mode(details: &MemWalIndexDetails) -> Result<LsmMode> {
.and_then(|s| s.parse::<u32>().ok())
.filter(|n| *n > 0)
.ok_or_else(|| Error::Runtime {
message: "merge_insert: MemWAL bucket spec has a missing or invalid num_buckets parameter".to_string(),
message: "merge_insert: RTE bucket spec has a missing or invalid num_buckets parameter".to_string(),
})?;
Ok(LsmMode::Bucket { spec })
}
@@ -697,7 +697,7 @@ fn resolve_lsm_mode(details: &MemWalIndexDetails) -> Result<LsmMode> {
// LSM merge_insert execution
// =============================================================================
/// Execute a `merge_insert` through the MemWAL shard writer cache.
/// Execute a `merge_insert` through the RTE shard writer cache.
///
/// The entire input is collected, schema-aligned, and shard-validated before
/// anything is written, then issued as a single atomic `ShardWriter::put` — so
@@ -716,7 +716,7 @@ pub(crate) async fn execute_lsm_merge_insert(
// Collect, align and shard-validate the whole input before writing
// anything. `ShardWriter::put` is atomic over the batch vector, so any
// failure raised here leaves the MemWAL untouched.
// failure raised here leaves the RTE untouched.
let mut batches: Vec<RecordBatch> = Vec::new();
let mut total_rows: u64 = 0;
@@ -803,7 +803,7 @@ fn resolve_batch_shard(
.downcast_ref::<Int32Array>()
.ok_or_else(|| Error::Runtime {
message: format!(
"merge_insert: MemWAL bucket evaluator returned {:?}; expected Int32",
"merge_insert: RTE bucket evaluator returned {:?}; expected Int32",
values.data_type()
),
})?;
@@ -824,7 +824,7 @@ fn resolve_batch_shard(
Ok(bucket_shard_id(u32::try_from(first).map_err(|_| {
Error::Runtime {
message: format!(
"merge_insert: MemWAL bucket evaluator returned negative bucket {}",
"merge_insert: RTE bucket evaluator returned negative bucket {}",
first
),
}
@@ -856,7 +856,7 @@ fn evaluate_lsm_shard_values(
if values.num_columns() != 1 {
return Err(Error::Runtime {
message: format!(
"merge_insert: MemWAL sharding spec evaluated to {} fields; expected exactly one",
"merge_insert: RTE sharding spec evaluated to {} fields; expected exactly one",
values.num_columns()
),
});
@@ -996,7 +996,7 @@ fn shard_writer_config_from_defaults(defaults: &HashMap<String, String>) -> Shar
}
/// Re-attach the dataset's Arrow schema (including field metadata) to a
/// user-supplied input batch. The MemWAL `ShardWriter` checks batch schemas
/// user-supplied input batch. The RTE `ShardWriter` checks batch schemas
/// against the dataset schema by exact equality, so input readers built
/// without the primary-key metadata must be rewrapped before being put.
///
rust/lancedb/src/table/query.rs
+19 -19
View File
@@ -94,9 +94,9 @@ async fn can_execute_namespace_query(table: &NativeTable, query: &AnyQuery) -> R
{
return Ok(false);
}
// A MemWAL write spec means reads auto-route through the LSM scanner in
// `create_plan` even when `use_memwal` is unset. The namespace request has no
// use_memwal field, so pushing the default query down would silently omit
// An RTE write spec means reads auto-route through the LSM scanner in
// `create_plan` even when `use_rte` is unset. The namespace request has no
// use_rte field, so pushing the default query down would silently omit
// un-compacted rows — force local execution whenever a spec is installed.
let dataset = table.dataset.get().await?;
if dataset.mem_wal_index_details().await?.is_some() {
@@ -106,10 +106,10 @@ async fn can_execute_namespace_query(table: &NativeTable, query: &AnyQuery) -> R
}
fn requires_local_namespace_execution(query: &AnyQuery) -> bool {
// The namespace QueryTable request cannot represent approx_mode, search_effort or use_memwal.
// For use_memwal that is worse than a tuning miss: MemWAL read routing lives only
// The namespace QueryTable request cannot represent approx_mode, search_effort or use_rte.
// For use_rte that is worse than a tuning miss: RTE read routing lives only
// in `create_plan`, so a pushed-down query would return stale base-only data.
if query.base().use_memwal.is_some() || query.base().take_offsets.is_some() {
if query.base().use_rte.is_some() || query.base().take_offsets.is_some() {
return true;
}
matches!(
@@ -167,25 +167,25 @@ pub async fn create_plan(
let ds_ref = table.dataset.get().await?;
// MemWAL read routing driven by `use_memwal`:
// RTE read routing driven by `use_rte`:
// * unset — route through the LSM scanner iff the table carries a write spec
// * Some(true) — force MemWAL routing; error if the table has no write spec
// * Some(false) — read the base table only, bypassing the MemWAL
// * Some(true) — force RTE routing; error if the table has no write spec
// * Some(false) — read the base table only, bypassing the RTE
// The LSM scanner surfaces in-flight `merge_insert` data (active/frozen
// memtables + SSTables); validation and dispatch live in `lsm`.
let has_spec = ds_ref.mem_wal_index_details().await?.is_some();
let use_memwal = match query.base.use_memwal {
let use_rte = match query.base.use_rte {
Some(true) if !has_spec => {
return Err(Error::InvalidInput {
message: "use_memwal(true) was set but the table has no MemWAL write spec; \
install one with enable_memwal or leave use_memwal unset"
message: "use_rte(true) was set but the table has no RTE write spec; \
install one with enable_rte or leave use_rte unset"
.to_string(),
});
}
Some(enable) => enable,
None => has_spec,
};
if use_memwal {
if use_rte {
return lsm::create_lsm_plan(table, ds_ref, query).await;
}
@@ -1195,7 +1195,7 @@ mod tests {
table.set_unenforced_primary_key(["id"]).await.unwrap();
table
.enable_memwal(crate::table::MemWalWriteSpec::unsharded())
.enable_rte(crate::table::RTEWriteSpec::unsharded())
.await
.unwrap();
let lsm_error = table
@@ -1395,7 +1395,7 @@ mod tests {
}
#[tokio::test]
async fn test_execute_query_use_memwal_with_namespace_pushdown_runs_locally() {
async fn test_execute_query_use_rte_with_namespace_pushdown_runs_locally() {
use crate::connect;
use crate::table::query::execute_query;
use arrow_array::{Int32Array, RecordBatch};
@@ -1418,7 +1418,7 @@ mod tests {
.unwrap();
let table = conn
.create_table("test_use_memwal_namespace_fallback", batch)
.create_table("test_use_rte_namespace_fallback", batch)
.execute()
.await
.unwrap();
@@ -1429,13 +1429,13 @@ mod tests {
.pushdown_operations
.insert(NamespaceClientPushdownOperation::QueryTable);
// `use_memwal` set (even to false) must force local execution — the namespace
// request has no use_memwal field, so a pushdown would silently ignore it.
// `use_rte` set (even to false) must force local execution — the namespace
// request has no use_rte field, so a pushdown would silently ignore it.
let query_vector = Arc::new(Float32Array::from(vec![0.0, 0.0]));
let query = AnyQuery::VectorQuery(VectorQueryRequest {
base: QueryRequest {
limit: Some(1),
use_memwal: Some(false),
use_rte: Some(false),
..Default::default()
},
column: Some("vector".to_string()),
rust/lancedb/src/table/query/lsm.rs
+25 -25
View File
@@ -1,12 +1,12 @@
// SPDX-License-Identifier: Apache-2.0
// SPDX-FileCopyrightText: Copyright The LanceDB Authors
//! MemWAL read path.
//! RTE read path.
//!
//! When a table has a MemWAL write spec installed (see [`enable_memwal`]),
//! When a table has an RTE write spec installed (see [`enable_rte`]),
//! reads are routed through Lance's [`LsmScanner`] instead of the plain
//! base-table scan unless the query sets
//! [`use_memwal(false)`](crate::query::QueryBase::use_memwal). This makes data
//! [`use_rte(false)`](crate::query::QueryBase::use_rte). This makes data
//! written via the LSM `merge_insert` path — which lives in the active/frozen
//! in-memory memtables and the SSTables until an external
//! compaction merges it into the base table — visible to queries, deduplicated by
@@ -17,10 +17,10 @@
//! three run through a single [`LsmScanner`], so a `where` filter is honored as a
//! prefilter uniformly — including for vector search, where `LsmScanner` threads
//! it into the vector planner's prefilter. Shapes the LSM path cannot honor are
//! rejected with [`Error::NotSupported`]; the caller must set `use_memwal(false)` to
//! rejected with [`Error::NotSupported`]; the caller must set `use_rte(false)` to
//! run those against the base table.
//!
//! [`enable_memwal`]: crate::Table::enable_memwal
//! [`enable_rte`]: crate::Table::enable_rte
use std::collections::HashMap;
use std::sync::Arc;
@@ -52,12 +52,12 @@ use crate::utils::default_vector_column;
/// widen the per-source fetch to keep result pages filled.
const LSM_OVERFETCH_FACTOR: f64 = 2.0;
/// Build the LSM read plan for a MemWAL-routed query.
/// Build the LSM read plan for an RTE-routed query.
///
/// The caller guarantees `ds_ref` carries a MemWAL write spec (routing is decided
/// The caller guarantees `ds_ref` carries an RTE write spec (routing is decided
/// in [`create_plan`](super::create_plan)). Errors with [`Error::NotSupported`]
/// for query shapes the LSM scanner cannot honor — the caller must set
/// `use_memwal(false)` to run those against the base table.
/// `use_rte(false)` to run those against the base table.
pub(super) async fn create_lsm_plan(
table: &NativeTable,
ds_ref: Arc<Dataset>,
@@ -67,11 +67,11 @@ pub(super) async fn create_lsm_plan(
// A time-traveled (checked-out) handle pins an older dataset version, but the
// WAL manifests and cached writer expose current live state — mixing them would
// surface WAL rows written after the requested version. `use_memwal(false)` reads
// surface WAL rows written after the requested version. `use_rte(false)` reads
// the base table at the pinned version.
if table.dataset.time_travel_version().is_some() {
return Err(Error::NotSupported {
message: "the MemWAL LSM scanner cannot read from a time-traveled dataset version; set use_memwal(false) to read the base table at this version".to_string(),
message: "the MemWAL LSM scanner cannot read from a time-traveled dataset version; set use_rte(false) to read the base table at this version".to_string(),
});
}
@@ -80,7 +80,7 @@ pub(super) async fn create_lsm_plan(
.mem_wal_index_details()
.await?
.ok_or_else(|| Error::Runtime {
message: "the MemWAL write spec disappeared during read planning".to_string(),
message: "the RTE write spec disappeared during read planning".to_string(),
})?;
let pk_columns = pk_columns(&ds_ref)?;
@@ -135,12 +135,12 @@ pub(super) async fn create_lsm_plan(
restore_projection(plan, &query, &pk_columns)
}
/// Reject query shapes the MemWAL read path does not implement. On a MemWAL table
/// Reject query shapes the RTE read path does not implement. On an RTE table
/// reads route through the LSM scanner by default, so an unsupported shape is a
/// hard error rather than a silent fallback to the base-only scan — which would
/// exclude un-compacted MemWAL data. The caller must set `use_memwal(false)` to
/// exclude un-compacted RTE data. The caller must set `use_rte(false)` to
/// run these against the base table, accepting that the results omit un-compacted
/// MemWAL data.
/// RTE data.
///
/// A `where` filter is intentionally *not* rejected: every arm routes through
/// [`LsmScanner`], which applies it as a prefilter (see [`base_scanner`]).
@@ -148,7 +148,7 @@ fn reject_unsupported(query: &VectorQueryRequest) -> Result<()> {
let unsupported = |what: &str| {
Err(Error::NotSupported {
message: format!(
"the MemWAL LSM scanner does not support {what}; set use_memwal(false) to read the base table only (results will exclude un-compacted MemWAL data)"
"the MemWAL LSM scanner does not support {what}; set use_rte(false) to read the base table only (results will exclude un-compacted RTE data)"
),
})
};
@@ -413,7 +413,7 @@ fn base_scanner(
QueryFilter::Datafusion(expr) => scanner.filter_expr(expr.clone()),
QueryFilter::Substrait(_) => {
return Err(Error::NotSupported {
message: "the MemWAL LSM scanner does not support Substrait filters; set use_memwal(false) to read the base table only".to_string(),
message: "the MemWAL LSM scanner does not support Substrait filters; set use_rte(false) to read the base table only".to_string(),
});
}
};
@@ -458,11 +458,11 @@ async fn fts_plan(
offset: Option<usize>,
) -> Result<Arc<dyn ExecutionPlan>> {
// Pre-check for a lancedb-flavored error; `LsmScanner` also validates the
// single-column requirement, but without the `use_memwal(false)` guidance.
// single-column requirement, but without the `use_rte(false)` guidance.
let columns: Vec<String> = fts.columns().into_iter().collect();
if columns.len() > 1 {
return Err(Error::NotSupported {
message: "the MemWAL LSM scanner full-text search supports a single column; set use_memwal(false) to read the base table only".to_string(),
message: "the MemWAL LSM scanner full-text search supports a single column; set use_rte(false) to read the base table only".to_string(),
});
}
let column = columns.first().ok_or_else(|| Error::NotSupported {
@@ -482,7 +482,7 @@ async fn fts_plan(
{
return Err(Error::NotSupported {
message: format!(
"the MemWAL LSM scanner full-text search requires the FTS index on '{column}' to be maintained by the write spec (MemWalWriteSpec::with_maintained_indexes) and carried by every resident MemTable; otherwise un-compacted documents are omitted. set use_memwal(false) to read the base table only"
"the MemWAL LSM scanner full-text search requires the FTS index on '{column}' to be maintained by the write spec (RTEWriteSpec::with_maintained_indexes) and carried by every resident MemTable; otherwise un-compacted documents are omitted. set use_rte(false) to read the base table only"
),
});
}
@@ -515,7 +515,7 @@ fn resident_memtables_carry(
.all(|tables| carries(&tables.active) && tables.frozen.iter().all(&carries))
}
/// Whether an index of `type_url_suffix` covering `column` is in the MemWAL spec's
/// Whether an index of `type_url_suffix` covering `column` is in the RTE spec's
/// maintained set. Only a maintained index has its catch-up tracked (so exclusion is
/// gated correctly) and its in-memory arm kept current; an unmaintained base index
/// falls back to the compaction watermark and can drop rows it has not re-indexed.
@@ -633,7 +633,7 @@ fn resolve_single_index(
if names.len() > 1 {
return Err(Error::NotSupported {
message: format!(
"the MemWAL LSM scanner cannot resolve the {arm} index catch-up watermark for '{column}': it has multiple {arm} indexes; set use_memwal(false) to read the base table only"
"the MemWAL LSM scanner cannot resolve the {arm} index catch-up watermark for '{column}': it has multiple {arm} indexes; set use_rte(false) to read the base table only"
),
});
}
@@ -682,7 +682,7 @@ fn restore_projection(
///
/// Note: the base and SSTable arms use `fast_search` (indexed data only), so a
/// base-table row not covered by a vector index is invisible here — it surfaces
/// only via the memtable or `use_memwal(false)`.
/// only via the memtable or `use_rte(false)`.
#[allow(clippy::too_many_arguments)]
async fn vector_plan(
dataset: &Dataset,
@@ -717,7 +717,7 @@ async fn vector_plan(
if !index_maintained(dataset, &column, details, "VectorIndexDetails").await? {
return Err(Error::NotSupported {
message: format!(
"the MemWAL LSM scanner requires the vector index on '{column}' to be maintained by the write spec (MemWalWriteSpec::with_maintained_indexes); otherwise compacted rows not yet re-indexed are omitted. set use_memwal(false) to read the base table only"
"the MemWAL LSM scanner requires the vector index on '{column}' to be maintained by the write spec (RTEWriteSpec::with_maintained_indexes); otherwise compacted rows not yet re-indexed are omitted. set use_rte(false) to read the base table only"
),
});
}
@@ -726,7 +726,7 @@ async fn vector_plan(
// clear error rather than failing deep in the planner.
if is_binary_vector_column(&arrow_schema, &column) {
return Err(Error::NotSupported {
message: "the MemWAL LSM scanner does not support binary (uint8) vector search; set use_memwal(false) to read the base table only".to_string(),
message: "the MemWAL LSM scanner does not support binary (uint8) vector search; set use_rte(false) to read the base table only".to_string(),
});
}
@@ -813,7 +813,7 @@ mod tests {
};
assert!(
matches!(reject_unsupported(&query), Err(Error::NotSupported { message })
if message.contains("search_effort") && message.contains("un-compacted MemWAL"))
if message.contains("search_effort") && message.contains("un-compacted RTE"))
);
}
rust/lancedb/src/table/refresh.rs
+12 -18
View File
@@ -87,7 +87,7 @@ async fn execute_refresh_column_with_source(
column: &str,
) -> Result<RefreshExecution> {
table.dataset.ensure_mutable()?;
ensure_no_memwal_spec(table).await?;
ensure_no_rte_spec(table).await?;
let dataset = table.dataset.get().await?;
let expression = declared_expression(&dataset, column)?;
@@ -303,7 +303,7 @@ pub(crate) async fn execute_refresh_column_async(
// Validate before spawning so bad input is reported by this call rather
// than only by the job.
table.dataset.ensure_mutable()?;
ensure_no_memwal_spec(table).await?;
ensure_no_rte_spec(table).await?;
let dataset = table.dataset.get().await?;
declared_expression(&dataset, column)?;
drop(dataset);
@@ -323,12 +323,12 @@ pub(crate) async fn execute_refresh_column_async(
})))
}
/// Refuse to refresh under a MemWAL write spec.
/// Refuse to refresh under an RTE write spec.
///
/// Refresh enumerates base fragments, and a write spec keeps visible rows in
/// un-compacted MemWAL tiers it cannot reach -- success would silently omit
/// un-compacted RTE tiers it cannot reach -- success would silently omit
/// readable rows.
async fn ensure_no_memwal_spec(table: &NativeTable) -> Result<()> {
async fn ensure_no_rte_spec(table: &NativeTable) -> Result<()> {
use lance::dataset::mem_wal::DatasetMemWalExt;
// Unset drops the MemWAL index, so the spec alone stops describing a table
@@ -341,7 +341,7 @@ async fn ensure_no_memwal_spec(table: &NativeTable) -> Result<()> {
.list_mem_wal_latest_shard_ids()
.await?
.is_empty();
if retained_sstables || table.get_memwal_spec().await?.is_some() {
if retained_sstables || table.get_rte_spec().await?.is_some() {
return Err(Error::NotSupported {
message: "refresh_column is not supported on a table with an LSM write \
spec: rows in un-compacted tiers are invisible to refresh"
@@ -1637,7 +1637,7 @@ mod tests {
/// writer left in that state.
#[tokio::test]
async fn test_refresh_refuses_a_foreign_lsm_state() {
use crate::table::MemWalWriteSpec;
use crate::table::RTEWriteSpec;
let tmp_dir = tempfile::tempdir().unwrap();
let conn = connect(tmp_dir.path().to_str().unwrap())
@@ -1654,10 +1654,7 @@ mod tests {
.unwrap();
let table = conn.create_table("lsm", batch).execute().await.unwrap();
table.set_unenforced_primary_key(["x"]).await.unwrap();
table
.enable_memwal(MemWalWriteSpec::unsharded())
.await
.unwrap();
table.enable_rte(RTEWriteSpec::unsharded()).await.unwrap();
super::super::computed_columns::add_foreign_kind(&table, "doubled", "sql").await;
let err = table.refresh_column("doubled").await.unwrap_err();
@@ -1673,7 +1670,7 @@ mod tests {
/// flag still marks retained SSTable rows; refresh refuses on the flag.
#[tokio::test]
async fn test_refresh_refuses_retained_catchup_state() {
use crate::table::MemWalWriteSpec;
use crate::table::RTEWriteSpec;
let tmp_dir = tempfile::tempdir().unwrap();
let conn = connect(tmp_dir.path().to_str().unwrap())
@@ -1696,15 +1693,12 @@ mod tests {
.await
.unwrap();
table.set_unenforced_primary_key(["x"]).await.unwrap();
table
.enable_memwal(MemWalWriteSpec::unsharded())
.await
.unwrap();
table.enable_rte(RTEWriteSpec::unsharded()).await.unwrap();
let mut merge = table.merge_insert(&["x"]);
merge
.when_matched_update_all(None)
.when_not_matched_insert_all()
.use_memwal(true);
.use_rte(true);
merge
.execute(Box::new(arrow_array::RecordBatchIterator::new(
vec![Ok(batch)],
@@ -1712,7 +1706,7 @@ mod tests {
)))
.await
.unwrap();
table.disable_memwal().await.unwrap();
table.disable_rte().await.unwrap();
super::super::computed_columns::add_foreign_kind(&table, "doubled", "sql").await;
let err = table.refresh_column("doubled").await.unwrap_err();
rust/lancedb/src/table/memwal_stats.rs → rust/lancedb/src/table/rte_stats.rs
+7 -7
View File
@@ -1,7 +1,7 @@
// SPDX-License-Identifier: Apache-2.0
// SPDX-FileCopyrightText: Copyright The LanceDB Authors
//! Live per-table-shard MemWAL state — the shape [`crate::Table::get_memwal_stats`]
//! Live per-table-shard RTE state — the shape [`crate::Table::get_rte_stats`]
//! returns and [`super::checkpoint`] polls.
//!
//! Nothing here is derived: sums and differences (total SSTable bytes, WAL lag)
@@ -79,18 +79,18 @@ impl TableShardStats {
}
}
/// Live MemWAL state, one entry per table shard.
/// Live RTE state, one entry per table shard.
#[derive(Debug, Clone, Deserialize)]
pub struct MemWalStats {
pub struct RTEStats {
pub table_shards: Vec<TableShardStats>,
}
/// Server-side JSON envelope for `get_memwal_stats`. `memwal_stats` is null when
/// the table has no MemWAL write path.
/// Server-side JSON envelope for `get_rte_stats`. `rte_stats` is null when
/// the table has no RTE write path.
#[derive(Debug, Deserialize)]
pub(crate) struct GetMemWalStatsResponse {
pub(crate) struct GetRTEStatsResponse {
#[serde(default)]
pub memwal_stats: Option<MemWalStats>,
pub rte_stats: Option<RTEStats>,
}
#[cfg(test)]
rust/lancedb/src/table/schema_evolution.rs
+2 -2
View File
@@ -131,7 +131,7 @@ pub(crate) async fn execute_declare(
) -> Result<AddColumnsResult> {
use lance::dataset::mem_wal::DatasetMemWalExt;
// A MemWAL write spec keeps visible rows in tiers refresh cannot reach;
// An RTE write spec keeps visible rows in tiers refresh cannot reach;
// checked against latest committed state, not this handle's snapshot.
table.checkout_latest().await?;
computed_columns::ensure_not_function_bound(
@@ -149,7 +149,7 @@ pub(crate) async fn execute_declare(
.list_mem_wal_latest_shard_ids()
.await?
.is_empty();
if retained_sstables || table.get_memwal_spec().await?.is_some() {
if retained_sstables || table.get_rte_spec().await?.is_some() {
return Err(Error::NotSupported {
message: "computed columns are not supported on a table with an LSM write \
spec: rows in un-compacted tiers are invisible to refresh"
rust/lancedb/tests/blob_integration.rs
+3 -3
View File
@@ -1981,8 +1981,8 @@ async fn blob_layout_overwrite_replaces_schema_in_both_directions() -> Result<()
}
#[tokio::test]
async fn blob_layout_memwal_merge_normalizes_before_writing() -> Result<()> {
use lancedb::{query::Select, table::MemWalWriteSpec};
async fn blob_layout_rte_merge_normalizes_before_writing() -> Result<()> {
use lancedb::{query::Select, table::RTEWriteSpec};
for wide in [false, true] {
let tmp = tempdir().unwrap();
let db = connect(tmp.path().to_str().unwrap()).execute().await?;
@@ -1991,7 +1991,7 @@ async fn blob_layout_memwal_merge_normalizes_before_writing() -> Result<()> {
.execute()
.await?;
table.set_unenforced_primary_key(["id"]).await?;
table.enable_memwal(MemWalWriteSpec::unsharded()).await?;
table.enable_rte(RTEWriteSpec::unsharded()).await?;
let mut merge = table.merge_insert(&["id"]);
merge
.when_matched_update_all(None)