* fix(query): preserve global fetch through physical-optimizer repartitioning
Aggregate soft-limit pushdown inserts a global CoalescePartitionsExec(fetch)
below an AggregateExec whose input is HashPartitioned/KeyPartitioned. The
DataFusion physical optimizer's EnforceDistribution/EnforceSorting passes then
dropped or rewrote that fetched coalesce, losing the global limit. Root-caused
and fixed in the DataFusion fork (GreptimeTeam/datafusion PR #35); this repo
pins to that fix commit and adds regression coverage.
- pin datafusion to GreptimeTeam/datafusion 48510c7f5 (fetch preservation fix)
- global_limit.rs: adapt to new DF API (input_distribution_requirements,
child_distribution, replace_children+Recompute); accept HashPartitioned and
KeyPartitioned as partitioning to restore; add regression unit tests
- sqlness: extend standalone common ssts and refresh distributed ssts_limit
result for the preserved fetch plan
Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com>
* refactor: drop dead HashPartitioned compat arms and test wrapper
- remove HashPartitioned compatibility matches (pinned DF only emits
KeyPartitioned); drop their #[expect(deprecated)] attributes
- remove single-use agg_with_limit helper whose seed limit is always
overwritten by the soft-limit transform
- restore Cargo.lock dependency edges unrelated to the DataFusion pin
bump (cargo update --precise had downgraded 26 unrelated packages)
Addresses oracle ablation review.
Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com>
* chore(deps): pin DataFusion to merged fetch preservation fix
Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com>
---------
Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com>
* fix(query): keep count_values generated label in enclosing expressions
`count_values("v", m)` projects the generated label as a real output
column, but did not register it in `ctx.tag_columns`. Enclosing
expressions (abs/round/+1/topk/label_replace/vector join) rebuild their
projection from `ctx.tag_columns` and silently drop the label.
Register the generated label in `ctx.tag_columns` after the projection,
and give it the same qualifier as other tag columns so qualified
references resolve. PromQL overwrites an input label with the same name,
so drop any existing tag with that name first to avoid duplicate column
ambiguity.
Fixes https://github.com/GreptimeTeam/greptimedb/issues/9181
Report: .e-agent/greptimedb_promql_compatibility_report_2026-09-16.md P0-1
Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com>
* fix(query): use total_cmp in count_values test helper
Silence clippy::needless_borrow on partial_cmp(&right.1); f64 sorting
uses total_cmp, matching the other planner test helpers.
Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com>
* fix(flow): keep count_values generated label as sink table primary key
FindGroupByFinalName::f_up only renamed a group key when the projection
aliased the key column directly. count_values now projects its generated
label as a unary UDF over the sampled column
(prom_float_to_string(value) AS label) while the aggregate still groups
by the raw column, so the name match failed and the sink table lost the
label from its PRIMARY KEY, demoting it to a DOUBLE value column.
Allow a projection above the aggregate to rename a group key by deriving
its output from a single group-key column (unary scalar function or cast
over that column only). Multi-column expressions, case, aggregates,
windows, subqueries and literals are still rejected so a computed column
cannot be mistaken for the group key.
Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com>
* fix(query): group count_values by the formatted sample value
The numeric count_values branch grouped by the raw sample column and only
formatted the value into the generated label in the post-aggregate
projection. Two distinct raw values that collapse to one label text
(e.g. BIGINT 9007199254740992 and 9007199254740993, both 9007199254740992
in Float64) were split into two groups, each emitting the same label set
at one timestamp, violating Prometheus' unique-label-set-per-timestamp
invariant.
Return the formatted value expression (prom_float_to_string, with a
CAST to Float64 for non-Float64 inputs) as previous_field_expressions so
the aggregate groups by the same expression that produces the label,
mirroring the existing mixed float/native-histogram precedent.
Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com>
* fix(query): drop needless borrow in count_values test helper
Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com>
* fix(flow): don't replace group key with derived unary expr
FindGroupByFinalName treated any unary expression of a single group
column (scalar fn, Cast, TryCast) as a rename of that group key and
swapped it in as the sink primary key. A derived expression such as
lower(host) is many-to-one, so distinct groups (HOST_A vs host_a) would
collapse to the same primary key and be silently deduplicated.
Narrow the matching to direct name-matched aliases of the actual group
expression only, and drop the derived-unary path (is_unary_expr_of_column)
and the allow_derived flag. count_values already groups by the formatted
expression name, so it is unaffected.
Add a regression asserting host survives and host_lc does not replace it
under both optimizer settings.
Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com>
---------
Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com>
* fix(query): insert MergeScan into nested scalar subqueries
DataFusion 55 keeps uncorrelated scalar subqueries as expression
subqueries (enable_physical_uncorrelated_scalar_subquery, default true)
instead of decorrelating them into joins, and executes them via the new
physical ScalarSubqueryExec. DistPlannerAnalyzer::try_push_down walked
the plan with a plain TreeNode transform that does not descend into
expression subqueries, so MergeScan was only inserted for depth-1
subqueries. A scalar subquery nested inside another scalar subquery kept
a bare frontend DistTable TableScan and failed at execution with
"Unsupported operation: get stream from a distributed table".
Use the subquery-aware transform so handle_subquery (PlanRewriter /
MergeScan insertion) runs for subquery plans at every nesting depth.
Fixes#9260.
Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com>
* test(query): strengthen nested scalar subquery regression coverage
Address review on #9261:
- Replace the ineffective 'no bare TableScan' string check with a real
subquery-aware plan walk (apply_with_subqueries); MergeScan hides its
remote input from traversal, so any TableScan the walk reaches was
genuinely left unwrapped.
- Add a distributed regression case on a range-partitioned table so the
nested inner aggregate must merge partial results across regions
(global AVG feeding an outer SUM filter).
Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com>
---------
Signed-off-by: discord9 <55937128+discord9@users.noreply.github.com>
* feat: add sync region instruction for repartition procedure
This commit introduces a new sync region instruction and integrates it
into the repartition procedure flow, specifically for metric engine tables.
Changes:
- Add SyncRegion instruction type and SyncRegionsReply in instruction.rs
- Implement SyncRegionHandler in datanode to handle sync region requests
- Add SyncRegion state in repartition procedure to sync newly allocated regions
- Integrate sync region step after enter_staging_region for metric engine tables
- Add sync_region flag and allocated_region_ids to PersistentContext
- Make SyncRegionFromRequest serializable for instruction transmission
- Add test utilities and mock support for sync region operations
The sync region step is conditionally executed based on the table engine type,
ensuring that newly allocated regions in metric engine tables are properly
synced from their source regions before proceeding with manifest remapping.
Signed-off-by: WenyXu <wenymedia@gmail.com>
* chore: add logs
Signed-off-by: WenyXu <wenymedia@gmail.com>
* feat(repartition): improve staging region handling and support metric engine repartition
- Reorder sync region flow: move SyncRegion from EnterStagingRegion to RepartitionStart to sync before applying staging
- Add ExitStaging metadata update state to properly clear staging leader info after repartition completes
- Update build_template_from_raw_table_info to optionally skip metric engine internal columns when creating region requests
- Fix region state transition: set_dropping now expects specific state (Staging or Writable) for proper validation
- Adjust region drop and copy handlers to handle staging regions correctly
- Add comprehensive test cases for metric engine SPLIT/MERGE partition operations on physical tables with logical tables
- Improve logging for table route updates, region drops, and repartition operations
Signed-off-by: WenyXu <wenymedia@gmail.com>
* refactor: removes code duplication
Signed-off-by: WenyXu <wenymedia@gmail.com>
* fix: update result
Signed-off-by: WenyXu <wenymedia@gmail.com>
* chore: refine comments
Signed-off-by: WenyXu <wenymedia@gmail.com>
* feat: add error strategy support for flush region and flush pending deallocate regions
- **Add `ErrorStrategy` enum** in `procedure/utils.rs`:
- Supports `Ignore` and `Retry` strategies for error handling
- Refactor `flush_region` to accept `error_strategy` parameter
- Extract `handle_flush_region_reply` helper function for better code organization
- **Add pending deallocate region support**:
- Add `pending_deallocate_region_ids` field to `PersistentContext`
- Implement `flush_pending_deallocate_regions` in `EnterStagingRegion` state
- Flush pending deallocate regions before entering staging regions to ensure data consistency
- **Update error handling**:
- `flush_leader_region`: Use `ErrorStrategy::Ignore` to skip unreachable datanodes
- `sync_region`: Use `ErrorStrategy::Retry` for critical operations
- `enter_staging_region`: Use `ErrorStrategy::Retry` when flushing pending deallocate regions
This change improves the robustness of the repartition procedure by:
1. Providing flexible error handling strategies for flush operations
2. Ensuring pending deallocate regions are properly flushed before repartitioning
3. Preventing data inconsistency during region migration
Signed-off-by: WenyXu <wenymedia@gmail.com>
* chore: apply suggestions from CR
Signed-off-by: WenyXu <wenymedia@gmail.com>
* fix: compile
Signed-off-by: WenyXu <wenymedia@gmail.com>
---------
Signed-off-by: WenyXu <wenymedia@gmail.com>
* feat: add repartition procedure factory support to DdlManager
- Introduce RepartitionProcedureFactory trait for creating and registering
repartition procedures
- Implement DefaultRepartitionProcedureFactory for metasrv with full support
- Implement StandaloneRepartitionProcedureFactory for standalone (unsupported)
- Add procedure loader registration for RepartitionProcedure and
RepartitionGroupProcedure
- Add helper methods to TableMetadataAllocator for allocator access
- Add error types for repartition procedure operations
- Update DdlManager to accept and use RepartitionProcedureFactoryRef
Signed-off-by: WenyXu <wenymedia@gmail.com>
* feat: integrate repartition procedure into DdlManager
- Add submit_repartition_task() to handle repartition from alter table
- Route Repartition operations in submit_alter_table_task() to repartition factory
- Refactor: rename submit_procedure() to execute_procedure_and_wait()
- Make all DDL operations wait for completion by default
- Add submit_procedure() for fire-and-forget submissions
- Add CreateRepartitionProcedure error type
- Add placeholder Repartition handling in grpc-expr (unsupported)
- Update greptime-proto dependency
Signed-off-by: WenyXu <wenymedia@gmail.com>
* feat: implement ALTER TABLE REPARTITION procedure submission
Signed-off-by: WenyXu <wenymedia@gmail.com>
* refactor(repartition): handle central region in apply staging manifest
- Introduce ApplyStagingManifestInstructions struct to organize instructions
- Add special handling for central region when applying staging manifests
- Transition state from UpdateMetadata to RepartitionEnd after applying staging manifests
- Remove next_state() method in RepartitionStart and inline state transitions
- Improve logging and expression serialization in DDL statement executor
- Move repartition tests from standalone to distributed test suite
Signed-off-by: WenyXu <wenymedia@gmail.com>
* chore: apply suggestions from CR
Signed-off-by: WenyXu <wenymedia@gmail.com>
* chore: update proto
Signed-off-by: WenyXu <wenymedia@gmail.com>
---------
Signed-off-by: WenyXu <wenymedia@gmail.com>
* feat(mysql): add SHOW WARNINGS support and return warnings for unsupported SET variables
Signed-off-by: Dennis Zhuang <killme2008@gmail.com>
* feat(function): add MySQL IF() function and PostgreSQL description functions for connector compatibility
Signed-off-by: Dennis Zhuang <killme2008@gmail.com>
* fix: show tables for mysql
Signed-off-by: Dennis Zhuang <killme2008@gmail.com>
* fix: partitions table in information_schema and add starrocks external catalog compatibility
Signed-off-by: Dennis Zhuang <killme2008@gmail.com>
* refactor: async udf
Signed-off-by: Dennis Zhuang <killme2008@gmail.com>
* fix: set warnings
Signed-off-by: Dennis Zhuang <killme2008@gmail.com>
* feat: impl pg_my_temp_schema and make description functions simple
Signed-off-by: Dennis Zhuang <killme2008@gmail.com>
* test: add test for issue 7313
Signed-off-by: Dennis Zhuang <killme2008@gmail.com>
* feat: apply suggestions
Signed-off-by: Dennis Zhuang <killme2008@gmail.com>
* fix: partition_expression and partition_description
Signed-off-by: Dennis Zhuang <killme2008@gmail.com>
* fix: test
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
Signed-off-by: Dennis Zhuang <killme2008@gmail.com>
* fix: unit tests
Signed-off-by: Dennis Zhuang <killme2008@gmail.com>
* fix: saerch_path only works for pg
Signed-off-by: Dennis Zhuang <killme2008@gmail.com>
* feat: improve warnings processing
Signed-off-by: Dennis Zhuang <killme2008@gmail.com>
* fix: warnings while writing affected rows and refactor
Signed-off-by: Dennis Zhuang <killme2008@gmail.com>
* chore: improve ShobjDescriptionFunction signature
Signed-off-by: Dennis Zhuang <killme2008@gmail.com>
* refactor: array_to_boolean
Signed-off-by: Dennis Zhuang <killme2008@gmail.com>
---------
Signed-off-by: Dennis Zhuang <killme2008@gmail.com>
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
discord9
shuiyisong
Yingwen
jeremyhi
fys
Ruihang Xia
Weny Xu
discord9
QuakeWanganddennis zhuang
zyy17andYingwen