feat: add view CRUD APIs (#4236)

A view is a named query a database stores and plans on every read. It
holds no rows, which is the whole difference from a materialized view.

## API

| Verb | Route |
| --- | --- |
| `create_view(name, query, namespace_path)` | `POST
/v1/view/{id}/create` |
| `describe_view(name, namespace_path)` | `POST /v1/view/{id}/describe`
|
| `drop_view(name, namespace_path)` | `POST /v1/view/{id}/drop` |
| `list_views(namespace_path)` | `GET /v1/namespace/{id}/view/list` |

On `Connection` and the `Database` trait, with the remote client, Python
(sync and async) and Node bindings. Local databases return
`NotSupported`: the server side is Sophon's, where a view is an object
of the database manifest.

`ViewDescription` carries the defining query, the database *and
namespace path* unqualified names in it resolve against, and the schema
the query resolved to. `create_view` returns one, so a caller has the
schema without a second call.

Both defaults travel with the view because it outlives the session that
declared it: the server re-plans the stored query on every read, so a
reader resolving an unqualified name against its own defaults would read
a different table. `default_namespace_path` crosses the wire as
`default_namespace`, a path like `namespace`, absent for the root.

There is no replace: a name already taken is an error, and changing a
view is a drop followed by a create, each authorized against what it
actually touches.

Querying a view stays SQL's job. There are no rows behind a view, so
there is no `open_view` returning a `Table`.
This commit is contained in:
Jack Ye
2026-09-23 15:09:40 +08:00
committed by GitHub
parent dd2539c2ca
commit 8541d6d9df
22 changed files with 1283 additions and 4 deletions
+67
View File
@@ -31,6 +31,7 @@ import type {
ListNamespacesResponse,
ListTablesResponse,
} from "./native";
import { ViewDescription, viewDescriptionFromNative } from "./view";
export type {
CreateNamespaceResponse,
DescribeNamespaceResponse,
@@ -377,6 +378,45 @@ export abstract class Connection {
namespacePath?: string[],
): Promise<Job>;
/**
* Create a view: a named query the database plans on every read.
*
* The query is planned once, at creation, so one that cannot be planned is
* rejected now rather than at the first read. A view holds no rows, and its
* readers see its sources as they are at read time.
*
* There is no replace: a name already taken is an error, and changing a
* view is a drop followed by a create.
*/
abstract createView(
name: string,
query: string,
namespacePath?: string[],
): Promise<ViewDescription>;
/**
* What this database records about the view named `name`: its defining
* query and the schema that query resolved to.
*/
abstract describeView(
name: string,
namespacePath?: string[],
): Promise<ViewDescription>;
/**
* Drop the view named `name`.
*
* The tables it reads are untouched: a view holds no rows of its own.
*/
abstract dropView(name: string, namespacePath?: string[]): Promise<void>;
/**
* The names of the views in one namespace.
*
* Names only; a definition comes from {@link describeView}.
*/
abstract listViews(namespacePath?: string[]): Promise<string[]>;
abstract openTable(
name: string,
namespacePath?: string[],
@@ -696,6 +736,33 @@ export class LocalConnection extends Connection {
);
}
async createView(
name: string,
query: string,
namespacePath?: string[],
): Promise<ViewDescription> {
return viewDescriptionFromNative(
await this.inner.createView(name, query, namespacePath ?? []),
);
}
async describeView(
name: string,
namespacePath?: string[],
): Promise<ViewDescription> {
return viewDescriptionFromNative(
await this.inner.describeView(name, namespacePath ?? []),
);
}
async dropView(name: string, namespacePath?: string[]): Promise<void> {
return this.inner.dropView(name, namespacePath ?? []);
}
async listViews(namespacePath?: string[]): Promise<string[]> {
return this.inner.listViews(namespacePath ?? []);
}
async listTables(
namespacePathOrOptions?: string[] | Partial<ListTablesOptions>,
options?: Partial<ListTablesOptions>,
+1
View File
@@ -26,6 +26,7 @@ export {
MaterializedViewDefinition,
MaterializedViewSelect,
} from "./materialized_view";
export { ViewDescription } from "./view";
export { JsHeaderProvider as NativeJsHeaderProvider } from "./native.js";
// OpenTelemetry metrics bridge. Only the high-level entry point is public; the
+48
View File
@@ -0,0 +1,48 @@
// SPDX-License-Identifier: Apache-2.0
// SPDX-FileCopyrightText: Copyright The LanceDB Authors
import { Schema, tableFromIPC } from "apache-arrow";
import { ViewDescription as NativeViewDescription } from "./native";
/**
* What a database records about one view.
*
* A view holds no rows: it stores the statement that defines it and the
* schema that statement resolved to, so a reader sees its sources as they
* are at read time.
*/
export interface ViewDescription {
/** The view's name within its namespace. */
name: string;
/** The namespace holding the view; empty is the root namespace. */
namespacePath: string[];
/** The defining query, as the database stores it. */
query: string;
/** The database that unqualified table names in {@link query} resolve against. */
defaultDatabase: string;
/**
* The namespace path those unqualified names resolve against; empty is the
* root namespace.
*
* Recorded with the view because it outlives the session that declared it:
* a reader resolving the query against its own default namespace could read
* a different table than the view was defined over.
*/
defaultNamespacePath: string[];
/** The schema the defining query resolved to when the view was created. */
schema: Schema;
}
/** Decode the schema the binding hands over as an Arrow IPC file. */
export function viewDescriptionFromNative(
view: NativeViewDescription,
): ViewDescription {
return {
name: view.name,
namespacePath: view.namespacePath,
query: view.query,
defaultDatabase: view.defaultDatabase,
defaultNamespacePath: view.defaultNamespacePath,
schema: tableFromIPC(view.schema).schema,
};
}