Merge main into gatekeeper/fix-4393-1

nodejs/Cargo.toml

@@ -1,7 +1,7 @@
 [package]
 name = "lancedb-nodejs"
 edition.workspace = true
-version = "0.40.0-beta.12"
+version = "0.40.0-beta.13"
 publish = false
 license.workspace = true
 description.workspace = true

nodejs/__test__/arrow.test.ts

@@ -790,6 +790,43 @@ describe.each(arrowVersions)("Arrow %#", (arrow: ApacheArrow) => {
       }
     });
 
+    it("will allow null values around struct values with a schema", function () {
+      const schema = new Schema([
+        new Field("value", new Struct([new Field("nested", new Int32())])),
+      ]);
+      for (const { records, nullIndex } of [
+        {
+          records: [{ value: null }, { value: { nested: 2 } }],
+          nullIndex: 0,
+        },
+        {
+          records: [{ value: { nested: 1 } }, { value: null }],
+          nullIndex: 1,
+        },
+      ]) {
+        const table = makeArrowTable(records, { schema });
+        const values = table.getChild("value");
+
+        expect(values?.nullCount).toBe(1);
+        expect(values?.get(nullIndex)).toBeNull();
+      }
+    });
+
+    it("will reject non-object values for struct fields in a schema", function () {
+      const schema = new Schema([
+        new Field("value", new Struct([new Field("nested", new Int32())])),
+      ]);
+      for (const records of [
+        [{ value: 5 }],
+        [{ value: 5 }, { value: { nested: 1 } }],
+        [{ value: { nested: 1 } }, { value: "oops" }],
+      ]) {
+        expect(() => makeArrowTable(records, { schema })).toThrow(
+          /Expected a struct value for field value at row \d, got/,
+        );
+      }
+    });
+
     it("will allow a schema to be provided", async function () {
       await checkTableCreation(
         async (records, _, schema) =>

nodejs/__test__/blob.test.ts

@@ -1,9 +1,25 @@
 // SPDX-License-Identifier: Apache-2.0
 // SPDX-FileCopyrightText: Copyright The LanceDB Authors
 
-import { Field, Int64, List, Schema, Struct, Utf8 } from "apache-arrow";
-import { makeArrowTable } from "../lancedb/arrow";
+import {
+  Field,
+  Float,
+  Float32,
+  Int32,
+  Int64,
+  List,
+  Schema,
+  Struct,
+  Utf8,
+} from "apache-arrow";
+import {
+  convertToTable,
+  makeArrowTable,
+  resolveBlobInputs,
+} from "../lancedb/arrow";
 import { BlobFile, blob, coerceBlobValue, isBlobField } from "../lancedb/blob";
+import { EmbeddingFunction, LanceSchema } from "../lancedb/embedding";
+import { getRegistry } from "../lancedb/embedding/registry";
 
 describe("blob()", () => {
   it("marks the field as lance.blob.v2", () => {
@@ -46,6 +62,13 @@ describe("blob()", () => {
   });
 });
 
+class BlobStruct {
+  constructor(
+    public data?: unknown,
+    public uri?: unknown,
+  ) {}
+}
+
 describe("coerceBlobValue", () => {
   it.each([
     ["Buffer", Buffer.from("x"), { data: Buffer.from("x"), uri: null }],
@@ -65,6 +88,41 @@ describe("coerceBlobValue", () => {
       { uri: "s3://bucket/key" },
       { data: null, uri: "s3://bucket/key" },
     ],
+    [
+      "ArrayBuffer",
+      new Uint8Array([120]).buffer,
+      { data: new Uint8Array([120]), uri: null },
+    ],
+    [
+      "URL",
+      new URL("https://example.com/blob"),
+      { data: null, uri: "https://example.com/blob" },
+    ],
+    [
+      "data struct with ArrayBuffer",
+      { data: new Uint8Array([121]).buffer },
+      { data: new Uint8Array([121]), uri: null },
+    ],
+    [
+      "uri struct with URL",
+      { uri: new URL("file:///tmp/a.png") },
+      { data: null, uri: "file:///tmp/a.png" },
+    ],
+    [
+      "class instance with Uint8Array data",
+      new BlobStruct(new Uint8Array([122])),
+      { data: new Uint8Array([122]), uri: null },
+    ],
+    [
+      "class instance with ArrayBuffer data",
+      new BlobStruct(new Uint8Array([122]).buffer),
+      { data: new Uint8Array([122]), uri: null },
+    ],
+    [
+      "class instance with URL uri",
+      new BlobStruct(undefined, new URL("s3://bucket/key")),
+      { data: null, uri: "s3://bucket/key" },
+    ],
     ["null", null, null],
   ])("accepts %s", (_name, input, expected) => {
     expect(coerceBlobValue(input)).toEqual(expected);
@@ -76,7 +134,7 @@ describe("coerceBlobValue", () => {
     [
       "Int16Array",
       new Int16Array([1]),
-      /Blob data must be Buffer or Uint8Array/,
+      /Blob data must be Buffer, Uint8Array, ArrayBuffer, or Blob/,
     ],
     [
       "both data and uri",
@@ -88,11 +146,160 @@ describe("coerceBlobValue", () => {
       { data: null, uri: null },
       /exactly one of 'data' or 'uri'/,
     ],
+    ["Blob", new Blob(["x"]), /must be read asynchronously/],
+    ["data struct with Blob", { data: new Blob(["x"]) }, /asynchronously/],
+    [
+      "class instance with Blob data",
+      new BlobStruct(new Blob(["x"])),
+      /asynchronously/,
+    ],
+    [
+      "unknown struct field",
+      { data: Buffer.from("y"), mime: "image/png" },
+      /Blob struct values only support 'data' and 'uri', got 'mime'/,
+    ],
+    [
+      "misspelled struct field",
+      { uri: "s3://bucket/key", dat: Buffer.from("y") },
+      /got 'dat'/,
+    ],
+    [
+      "SharedArrayBuffer",
+      new SharedArrayBuffer(2),
+      /Blob column values must be Buffer, Uint8Array, ArrayBuffer, Blob/,
+    ],
+    ["Date", new Date(0), /Blob column values must be/],
   ])("rejects %s", (_name, input, message) => {
     expect(() => coerceBlobValue(input)).toThrow(message);
   });
 });
 
+describe("resolveBlobInputs", () => {
+  it("reads Blob and File values in blob columns only", async () => {
+    const schema = new Schema([
+      new Field("name", new Utf8(), true),
+      blob("image"),
+      new Field("images", new List(blob("image")), true),
+    ]);
+    const shared = new Blob(["same"]);
+    const notABlobColumn = new Blob(["other"]);
+    const data = [
+      {
+        name: notABlobColumn,
+        image: new File(["file"], "a.txt"),
+        images: [shared, { data: shared }, null],
+      },
+      { name: "x", image: Buffer.from("buf"), images: null },
+    ];
+    const resolved = await resolveBlobInputs(data, schema);
+
+    expect(resolved[0].name).toBe(notABlobColumn);
+    expect(Buffer.from(resolved[0].image as Uint8Array).toString()).toBe(
+      "file",
+    );
+    const images = resolved[0].images as [Uint8Array, { data: Uint8Array }];
+    expect(Buffer.from(images[0]).toString()).toBe("same");
+    expect(Buffer.from(images[1].data).toString()).toBe("same");
+    // Rows without Blob values and the caller's records are left alone.
+    expect(resolved[1]).toBe(data[1]);
+    expect(data[0].images?.[0]).toBe(shared);
+  });
+
+  it("reads each distinct Blob once, bare or inside data structs", async () => {
+    class CountingBlob extends Blob {
+      calls = 0;
+      async arrayBuffer(): Promise<ArrayBuffer> {
+        this.calls++;
+        return super.arrayBuffer();
+      }
+    }
+    const shared = new CountingBlob(["same"]);
+    const schema = new Schema([blob("image")]);
+    const resolved = await resolveBlobInputs(
+      [
+        { image: { data: shared } },
+        { image: { data: shared } },
+        { image: shared },
+      ],
+      schema,
+    );
+    expect(shared.calls).toBe(1);
+    const first = (resolved[0].image as { data: Uint8Array }).data;
+    expect(Buffer.from(first).toString()).toBe("same");
+    expect((resolved[1].image as { data: Uint8Array }).data).toBe(first);
+    expect(resolved[2].image).toBe(first);
+  });
+
+  it("reads a Blob held by a class instance", async () => {
+    const schema = new Schema([blob("image")]);
+    const [row] = await resolveBlobInputs(
+      [{ image: new BlobStruct(new Blob(["hi"])) }],
+      schema,
+    );
+    expect(
+      Buffer.from((row.image as { data: Uint8Array }).data).toString(),
+    ).toBe("hi");
+  });
+
+  it("returns the input when there is nothing to read", async () => {
+    const data = [{ image: Buffer.from("x") }];
+    const schema = new Schema([blob("image")]);
+    await expect(resolveBlobInputs(data, schema)).resolves.toBe(data);
+    await expect(resolveBlobInputs(data)).resolves.toBe(data);
+  });
+});
+
+describe("convertToTable blob columns with embedding functions", () => {
+  // Embedding metadata routes rows through ensureNestedFieldsExist, which must
+  // leave blob values for the blob coercion instead of treating them as structs.
+  class LengthEmbedding extends EmbeddingFunction<string> {
+    ndims() {
+      return 2;
+    }
+    embeddingDataType(): Float {
+      return new Float32();
+    }
+    async computeSourceEmbeddings(data: string[]) {
+      return data.map((text) => [text.length, 1]);
+    }
+  }
+  getRegistry().register("blob-test-length")(LengthEmbedding);
+  const func = getRegistry().get<LengthEmbedding>("blob-test-length")!.create();
+  const embeddingSchema = LanceSchema({
+    text: func.sourceField(new Utf8()),
+    vector: func.vectorField(),
+  });
+  const schema = new Schema(
+    [...embeddingSchema.fields, blob("image")],
+    embeddingSchema.metadata,
+  );
+
+  it.each([
+    ["Uint8Array", new Uint8Array([104, 105]), "hi", null],
+    ["ArrayBuffer", new Uint8Array([104, 105]).buffer, "hi", null],
+    ["Blob", new Blob(["hi"]), "hi", null],
+    ["URI string", "s3://b/k", null, "s3://b/k"],
+    ["URL", new URL("s3://b/k"), null, "s3://b/k"],
+    ["data struct", { data: new Uint8Array([104, 105]) }, "hi", null],
+  ])("accepts %s input", async (_name, image, data, uri) => {
+    const table = await convertToTable([{ text: "a", image }], undefined, {
+      schema,
+    });
+    const column = table.getChild("image")!;
+    expect(column.nullCount).toBe(0);
+    const bytes = column.getChild("data")!.get(0);
+    expect(bytes == null ? null : Buffer.from(bytes).toString()).toBe(data);
+    expect(column.getChild("uri")!.get(0)).toBe(uri);
+  });
+
+  it("writes a missing blob as null", async () => {
+    const table = await convertToTable([{ text: "a" }], undefined, { schema });
+    const column = table.getChild("image")!;
+    expect(column.nullCount).toBe(1);
+    expect(column.get(0)).toBeNull();
+  });
+});
+
 describe("BlobFile", () => {
   it("rejects constructing BlobFile without a native handle", () => {
     expect(() => new (BlobFile as unknown as { new (): BlobFile })()).toThrow(
@@ -103,26 +310,122 @@ describe("BlobFile", () => {
 
 describe("makeArrowTable blob columns", () => {
   it.each([
-    ["ArrayBuffer", new Uint8Array([104]).buffer],
-    ["Blob", new Blob(["hello"])],
-    ["File", new File(["hello"], "hello.txt")],
-    ["URL", new URL("https://example.com/blob")],
     ["ReadableStream", new ReadableStream()],
-    ["data struct with ArrayBuffer", { data: new Uint8Array([104]).buffer }],
-    ["uri struct with URL", { uri: new URL("https://example.com/blob") }],
+    ["Int16Array", new Int16Array([1])],
   ])("rejects %s input rather than writing a null blob", (_name, image) => {
     const schema = new Schema([blob("image")]);
     expect(() => makeArrowTable([{ image }], { schema })).toThrow(/image/);
   });
 
-  it("rejects Blob instances with enumerable fields as one invalid value", () => {
+  it.each([
+    ["Blob", new Blob(["hello"])],
+    ["File", new File(["hello"], "hello.txt")],
+    ["data struct with Blob", { data: new Blob(["hello"]) }],
+    [
+      "Blob with enumerable fields",
+      Object.assign(new Blob(["hello"]), { position: 0 }),
+    ],
+  ])("points %s input at the async APIs", (_name, image) => {
     const schema = new Schema([blob("image")]);
-    const image = Object.assign(new Blob(["hello"]), { position: 0 });
-    expect(() => makeArrowTable([{ image }], { schema })).toThrow(
-      "Unsupported object value for field image at row 0.",
+    expect(() => makeArrowTable([{ id: 1, image }], { schema })).toThrow(
+      /blob field image at row 0: Blob and File values must be read asynchronously/,
     );
   });
 
+  it.each([
+    ["ArrayBuffer", new Uint8Array([104, 105]).buffer, "hi", null],
+    [
+      "data struct with ArrayBuffer",
+      { data: new Uint8Array([104, 105]).buffer },
+      "hi",
+      null,
+    ],
+    ["URL", new URL("https://example.com/a"), null, "https://example.com/a"],
+    [
+      "uri struct with URL",
+      { uri: new URL("file:///tmp/a.png") },
+      null,
+      "file:///tmp/a.png",
+    ],
+  ])("accepts %s input", (_name, image, data, uri) => {
+    const schema = new Schema([blob("image")]);
+    const input = { image };
+    const table = makeArrowTable([input], { schema });
+    const column = table.getChild("image")!;
+    expect(column.nullCount).toBe(0);
+    const bytes = column.getChild("data")!.get(0);
+    expect(bytes == null ? null : Buffer.from(bytes).toString()).toBe(data);
+    expect(column.getChild("uri")!.get(0)).toBe(uri);
+    // The caller's record is left untouched.
+    expect(input.image).toBe(image);
+  });
+
+  it.each([
+    ["Uint8Array", new Uint8Array([104, 105]), "hi", null],
+    ["URI string", "s3://b/k", null, "s3://b/k"],
+    ["data struct", { data: new Uint8Array([104, 105]) }, "hi", null],
+  ])("accepts %s input after a null row", (_name, image, data, uri) => {
+    const schema = new Schema([blob("image")]);
+    const table = makeArrowTable([{ image: null }, { image }], { schema });
+    const column = table.getChild("image")!;
+    expect(column.nullCount).toBe(1);
+    expect(column.get(0)).toBeNull();
+    const bytes = column.getChild("data")!.get(1);
+    expect(bytes == null ? null : Buffer.from(bytes).toString()).toBe(data);
+    expect(column.getChild("uri")!.get(1)).toBe(uri);
+  });
+
+  it("accepts a null struct before one holding a blob", () => {
+    const schema = new Schema([
+      new Field("info", new Struct([blob("image")]), true),
+    ]);
+    const table = makeArrowTable(
+      [{ info: null }, { info: { image: new Uint8Array([97]) } }],
+      { schema },
+    );
+    const info = table.getChild("info")!;
+    expect(info.get(0)).toBeNull();
+    const image = (info.get(1) as { image: { data: Uint8Array } }).image;
+    expect(Buffer.from(image.data).toString()).toBe("a");
+  });
+
+  it("skips a missing blob column named like an Object.prototype member", () => {
+    const schema = new Schema([
+      new Field("id", new Int32(), true),
+      blob("constructor"),
+    ]);
+    const table = makeArrowTable([{ id: 1 }], { schema });
+    expect(table.schema.fields.map((f) => f.name)).toEqual(["id"]);
+  });
+
+  it("accepts ArrayBuffer and URL inside lists and structs", () => {
+    const schema = new Schema([
+      new Field("images", new List(blob("image")), true),
+      new Field("info", new Struct([blob("image")]), true),
+    ]);
+    const table = makeArrowTable(
+      [
+        {
+          images: [new Uint8Array([97]).buffer, new URL("s3://b/k")],
+          info: { image: new Uint8Array([98]).buffer },
+        },
+      ],
+      { schema },
+    );
+    const images = Array.from(
+      table.getChild("images")!.get(0) as Iterable<{
+        data: Uint8Array | null;
+        uri: string | null;
+      }>,
+    );
+    expect(Buffer.from(images[0].data!).toString()).toBe("a");
+    expect(images[1].uri).toBe("s3://b/k");
+    const info = table.getChild("info")!.get(0) as {
+      image: { data: Uint8Array };
+    };
+    expect(Buffer.from(info.image.data).toString()).toBe("b");
+  });
+
   it("coerces Buffer input onto a blob field", () => {
     const schema = new Schema([
       new Field("id", new Int64(), true),

nodejs/__test__/catalog.test.ts

@@ -80,9 +80,12 @@ describe("remote catalog", () => {
         const second = await catalog.connectDatabase("other");
         expect(await second.tableNames()).toEqual([]);
         expect(await first.tableNames()).toEqual([]);
-        expect(
-          await catalog.listDatabases({ limit: 1, pageToken: "a/b" }),
-        ).toEqual({ databases: ["team/search"], pageToken: "next" });
+        const names = catalog.listDatabases({ pageLimit: 1, pageToken: "a/b" });
+        expect(await names.next()).toEqual({
+          done: false,
+          value: "team/search",
+        });
+        expect(names.pageToken()).toBe("next");
         await catalog.dropDatabase("team/search", { ignoreMissing: true });
         expect(requests[0].url).toBe("/v1/namespace/team%2Fsearch/create");
         expect(requests[0].body).toEqual({ mode: "ExistOk" });
@@ -129,11 +132,139 @@ describe("remote catalog", () => {
     );
   });
 
+  it("lazily traverses pages, including empty pages, and exposes cached state", async () => {
+    await withCatalog(
+      [
+        // biome-ignore lint/style/useNamingConvention: server wire format
+        [200, { namespaces: ["a", "b"], page_token: "empty" }],
+        // biome-ignore lint/style/useNamingConvention: server wire format
+        [200, { namespaces: [], page_token: "last" }],
+        // biome-ignore lint/style/useNamingConvention: server wire format
+        [200, { namespaces: ["c", "d"], page_token: "" }],
+      ],
+      async (catalog, requests) => {
+        const names = catalog.listDatabases({ pageLimit: 2 });
+        expect(names[Symbol.asyncIterator]()).toBe(names);
+        expect(requests).toHaveLength(0);
+        expect(names.numPageResults()).toBe(0);
+        expect(names.pageToken()).toBeUndefined();
+        expect(await names.next()).toEqual({ done: false, value: "a" });
+        expect(names.numPageResults()).toBe(1);
+        expect(names.pageToken()).toBe("empty");
+        expect(await names.next()).toEqual({ done: false, value: "b" });
+        expect(requests).toHaveLength(1);
+        expect(await names.next()).toEqual({ done: false, value: "c" });
+        expect(requests).toHaveLength(3);
+        expect(names.pageToken()).toBeUndefined();
+        expect(names.numPageResults()).toBe(1);
+        const remaining = [];
+        for await (const name of names) remaining.push(name);
+        expect(remaining).toEqual(["d"]);
+        expect(names.numPageResults()).toBe(0);
+        expect(await names.next()).toEqual({ done: true, value: undefined });
+        expect(requests.map((request) => request.url)).toEqual([
+          "/v1/namespace/$/list?limit=2",
+          "/v1/namespace/$/list?limit=2&page_token=empty",
+          "/v1/namespace/$/list?limit=2&page_token=last",
+        ]);
+      },
+    );
+  });
+
+  it("resumes from a saved token after draining the cache", async () => {
+    await withCatalog(
+      [
+        // biome-ignore lint/style/useNamingConvention: server wire format
+        [200, { namespaces: ["a", "b"], page_token: "next/token" }],
+        [200, { namespaces: ["c"] }],
+      ],
+      async (catalog, requests) => {
+        const names = catalog.listDatabases({
+          pageToken: "start/token",
+          pageLimit: 2,
+        });
+        expect(names.pageToken()).toBe("start/token");
+        expect(await names.next()).toEqual({ done: false, value: "a" });
+        while (names.numPageResults() > 0) await names.next();
+        const resumed = catalog.listDatabases({
+          pageToken: names.pageToken(),
+          pageLimit: 2,
+        });
+        const results = [];
+        for await (const name of resumed) results.push(name);
+        expect(results).toEqual(["c"]);
+        expect(requests.map((request) => request.url)).toEqual([
+          "/v1/namespace/$/list?limit=2&page_token=start%2Ftoken",
+          "/v1/namespace/$/list?limit=2&page_token=next%2Ftoken",
+        ]);
+      },
+    );
+  });
+
+  it("terminates on errors and retains the failed request token", async () => {
+    await withCatalog(
+      [
+        // biome-ignore lint/style/useNamingConvention: server wire format
+        [200, { namespaces: ["a"], page_token: "retry" }],
+        [400, {}],
+        [200, { namespaces: ["b"] }],
+      ],
+      async (catalog, requests) => {
+        const names = catalog.listDatabases();
+        await names.next();
+        await expect(names.next()).rejects.toThrow();
+        expect(names.pageToken()).toBe("retry");
+        expect(names.numPageResults()).toBe(0);
+        expect(await names.next()).toEqual({ done: true, value: undefined });
+        expect(requests).toHaveLength(2);
+        const resumed = catalog.listDatabases({ pageToken: names.pageToken() });
+        expect(await resumed.next()).toEqual({ done: false, value: "b" });
+        expect(await resumed.next()).toEqual({ done: true, value: undefined });
+      },
+    );
+  });
+
+  it("handles an empty listing and concurrent advances", async () => {
+    await withCatalog(
+      [
+        [200, { namespaces: [] }],
+        [200, { namespaces: ["a", "b"] }],
+      ],
+      async (catalog, requests) => {
+        const empty = catalog.listDatabases();
+        expect(await empty.next()).toEqual({ done: true, value: undefined });
+        expect(await empty.next()).toEqual({ done: true, value: undefined });
+        const names = catalog.listDatabases();
+        const pending = [names.next(), names.next(), names.next()];
+        expect(names.numPageResults()).toBe(0);
+        expect(names.pageToken()).toBeUndefined();
+        const results = await Promise.all(pending);
+        expect(
+          results
+            .filter((result) => !result.done)
+            .map((result) => result.value)
+            .sort(),
+        ).toEqual(["a", "b"]);
+        expect(results.filter((result) => result.done)).toHaveLength(1);
+        expect(requests).toHaveLength(2);
+      },
+    );
+  });
+
   it("validates endpoints and pagination", async () => {
     await expect(connectCatalog("/tmp/catalog")).rejects.toThrow();
     const catalog = await connectCatalog("http://127.0.0.1:1");
-    for (const limit of [0, -1, 1.5, 2147483648]) {
-      await expect(catalog.listDatabases({ limit })).rejects.toThrow("limit");
+    for (const limit of [
+      0,
+      -1,
+      1.5,
+      2147483648,
+      Number.NaN,
+      Number.POSITIVE_INFINITY,
+    ]) {
+      expect(() => catalog.listDatabases({ pageLimit: limit })).toThrow(
+        "limit",
+      );
     }
     await expect(catalog.connectDatabase("a$b")).rejects.toThrow(
       "Invalid database name",

nodejs/__test__/materialized_view.test.ts

@@ -16,15 +16,11 @@ describe("materialized views", () => {
   beforeEach(async () => {
     tmpDir = tmp.dirSync({ unsafeCleanup: true });
     db = await connect(tmpDir.name);
-    await db.createTable(
-      "people",
-      [
-        { name: "ada", age: 36 },
-        { name: "kid", age: 7 },
-        { name: "grace", age: 85 },
-      ],
-      { storageOptions: { newTableEnableStableRowIds: "true" } },
-    );
+    await db.createTable("people", [
+      { name: "ada", age: 36 },
+      { name: "kid", age: 7 },
+      { name: "grace", age: 85 },
+    ]);
   });
   afterEach(() => tmpDir.removeCallback());
 
@@ -52,19 +48,9 @@ describe("materialized views", () => {
       "SELECT * FROM people",
     );
 
-    for (const [format, fn] of [
-      [3, "vector_duplicate_pairs"],
-      [4, "vector_dedup"],
-    ]) {
-      const native = `SELECT * FROM ${fn}('images', 3, 'phash', 4)`;
-      expect(read(JSON.stringify({ format, query: native })).query).toBe(
-        native,
-      );
-    }
-
     // A newer writer's layout is reported, never guessed at.
     for (const newer of [
-      `{"format":5,"query":${JSON.stringify(query)}}`,
+      `{"format":3,"query":${JSON.stringify(query)}}`,
       '{"kind":"select_v3","source_table":"people"}',
     ]) {
       expect(() => read(newer)).toThrow(/cannot refresh/);
@@ -106,7 +92,7 @@ describe("materialized views", () => {
     );
   });
 
-  it("refreshes incrementally after an append", async () => {
+  it("fully refreshes after an append", async () => {
     const view = await db.createMaterializedView("copy", "people", {
       withNoData: true,
     });
@@ -115,11 +101,11 @@ describe("materialized views", () => {
     const people = await db.openTable("people");
     await people.add([{ name: "alan", age: 41 }]);
     const result = await view.refresh();
-    expect(result.mode).toBe("incremental");
-    expect(Number(result.rowsWritten)).toBe(1);
+    expect(result.mode).toBe("rebuild");
+    expect(Number(result.rowsWritten)).toBe(4);
     expect(await view.table().countRows()).toBe(4);
 
-    expect((await view.refresh()).mode).toBe("no_op");
+    expect((await view.refresh()).mode).toBe("rebuild");
   });
 
   it("lists views and rejects non-views", async () => {
@@ -172,9 +158,7 @@ describe("materialized views", () => {
   });
 
   it("quotes bare select names", async () => {
-    await db.createTable("odd_names", [{ "order item": "widget" }], {
-      storageOptions: { newTableEnableStableRowIds: "true" },
-    });
+    await db.createTable("odd_names", [{ "order item": "widget" }]);
     const view = await db.createMaterializedView("quoted", "odd_names", {
       select: ["order item"],
       withNoData: true,
@@ -183,10 +167,9 @@ describe("materialized views", () => {
     expect(Number(result.rowsWritten)).toBe(1);
   });
 
-  it("requires stable row ids on the source", async () => {
+  it("does not require stable row ids on the source", async () => {
     await db.createTable("plain", [{ x: 1 }]);
-    await expect(db.createMaterializedView("v", "plain")).rejects.toThrow(
-      "stable row ids",
-    );
+    const view = await db.createMaterializedView("v", "plain");
+    expect(await view.table().countRows()).toBe(1);
   });
 });

nodejs/__test__/table.test.ts

@@ -9,6 +9,7 @@ import * as tmp from "tmp";
 
 import {
   AutoQuery,
+  type BlobInput,
   Connection,
   MatchQuery,
   PhraseQuery,
@@ -2583,7 +2584,7 @@ describe("when dealing with blob columns", () => {
       new Field("id", new Int64(), true),
       blob("image"),
     ]);
-    const row = { id: 1n, image: new Uint8Array([104]).buffer };
+    const row = { id: 1n, image: new ReadableStream() };
 
     await expect(db.createTable("invalid", [row], { schema })).rejects.toThrow(
       /field image at row 0/,
@@ -2613,6 +2614,8 @@ describe("when dealing with blob columns", () => {
   it.each([
     ["URI string", (uri: string) => uri],
     ["URI struct", (uri: string) => ({ uri })],
+    ["URL", (uri: string) => new URL(uri)],
+    ["URL struct", (uri: string) => ({ uri: new URL(uri) })],
   ])(
     "round-trips an external blob from a %s after reopening",
     async (_label, blobValue) => {
@@ -2706,6 +2709,63 @@ describe("when dealing with blob columns", () => {
     expect(bytes[0]).toEqual(payload);
   });
 
+  it("accepts ArrayBuffer, Blob, and File in createTable and add", async () => {
+    const db = await connect(tmpDir.name);
+    const schema = new Schema([
+      new Field("id", new Int64(), true),
+      blob("image"),
+    ]);
+    const rows: { id: bigint; image: BlobInput }[] = [
+      { id: 1n, image: new TextEncoder().encode("array-buffer").buffer },
+      { id: 2n, image: new Blob(["blob"]) },
+      { id: 3n, image: { data: new File(["file"], "f.txt") } },
+    ];
+    const table = await db.createTable("widened", rows.slice(0, 2), {
+      schema,
+    });
+    await table.add(rows.slice(2));
+
+    const results = await table.query().select(["id"]).withRowId().toArray();
+    results.sort((a, b) => Number(a.id - b.id));
+    const bytes = await table.fetchBlobs(
+      "image",
+      results.map((row) => row._rowid as bigint),
+    );
+    expect(bytes.map((b) => b?.toString())).toEqual([
+      "array-buffer",
+      "blob",
+      "file",
+    ]);
+  });
+
+  it("accepts Blob values in mergeInsert", async () => {
+    const db = await connect(tmpDir.name);
+    const schema = new Schema([
+      new Field("id", new Int64(), true),
+      blob("image"),
+    ]);
+    const table = await db.createTable(
+      "merge_blobs",
+      [{ id: 1n, image: Buffer.from("old") }],
+      { schema },
+    );
+    await table
+      .mergeInsert("id")
+      .whenMatchedUpdateAll()
+      .whenNotMatchedInsertAll()
+      .execute([
+        { id: 1n, image: new Blob(["new"]) },
+        { id: 2n, image: new Uint8Array([104, 105]).buffer },
+      ]);
+    const results = await table.query().select(["id"]).withRowId().toArray();
+    results.sort((a, b) => Number(a.id - b.id));
+    const bytes = await table.fetchBlobs(
+      "image",
+      results.map((row) => row._rowid as bigint),
+    );
+    expect(bytes.map((b) => b?.toString())).toEqual(["new", "hi"]);
+  });
+
   it("discovers blob columns", async () => {
     const { table } = await openBlobTable();
     expect(await table.blobColumns()).toEqual(["image"]);

nodejs/lancedb/arrow.ts

@@ -40,7 +40,12 @@ import {
 } from "apache-arrow";
 import { Buffers } from "apache-arrow/data";
 import { typedArrayToArrowType } from "./arrow_type";
-import { coerceBlobValue, isBlobField } from "./blob";
+import {
+  blobToRead,
+  coerceBlobValue,
+  isBlobField,
+  withBlobBytes,
+} from "./blob";
 import { type EmbeddingFunction } from "./embedding/embedding_function";
 import {
   EmbeddingFunctionConfig,
@@ -449,6 +454,24 @@ export function makeArrowTable(
     );
   }
 
+  if (schema !== undefined) {
+    // Validate blob values up front and give every one the full
+    // `{ data, uri }` shape, so inference sees the same struct whether a row
+    // passed bytes, a URI, or a partial struct.
+    data = mapBlobInputs(data, schema, (value, field, row) => {
+      if (value === undefined) {
+        return value;
+      }
+      try {
+        return coerceBlobValue(value);
+      } catch (e) {
+        throw new Error(
+          `Invalid value for blob field ${field} at row ${row}: ${(e as Error).message}`,
+        );
+      }
+    });
+  }
+
   let schemaMetadata = schema?.metadata || new Map<string, string>();
   if (metadata !== undefined) {
     schemaMetadata = new Map([...schemaMetadata, ...metadata]);
@@ -510,6 +533,133 @@ function containsBlobField(field: Field): boolean {
   );
 }
 
+type BlobInputVisitor = (value: unknown, field: string, row: number) => unknown;
+
+/**
+ * Calls `visit` on every value that lands in a blob column of `schema`,
+ * including blobs nested in structs and lists, and replaces it with the
+ * result. Records and arrays are copied only where a value changed, so the
+ * caller's data is never mutated.
+ */
+function mapBlobInputs(
+  data: Array<Record<string, unknown>>,
+  schema: Schema,
+  visit: BlobInputVisitor,
+): Array<Record<string, unknown>> {
+  if (!schema.fields.some(containsBlobField)) {
+    return data;
+  }
+  return data.map((record, row) =>
+    isObject(record)
+      ? mapBlobFields(record, schema.fields, "", row, visit)
+      : record,
+  );
+}
+
+function mapBlobFields(
+  record: Record<string, unknown>,
+  fields: Field[],
+  prefix: string,
+  row: number,
+  visit: BlobInputVisitor,
+): Record<string, unknown> {
+  let out: Record<string, unknown> | undefined;
+  for (const field of fields) {
+    if (!containsBlobField(field) || !Object.hasOwn(record, field.name)) {
+      continue;
+    }
+    const value = record[field.name];
+    const mapped = mapBlobField(
+      value,
+      field,
+      `${prefix}${field.name}`,
+      row,
+      visit,
+    );
+    if (mapped !== value) {
+      out ??= { ...record };
+      out[field.name] = mapped;
+    }
+  }
+  return out ?? record;
+}
+
+function mapBlobField(
+  value: unknown,
+  field: Field,
+  label: string,
+  row: number,
+  visit: BlobInputVisitor,
+): unknown {
+  if (isBlobField(field)) {
+    return visit(value, label, row);
+  }
+  if (field.type instanceof Struct && isObject(value)) {
+    return mapBlobFields(value, field.type.children, `${label}.`, row, visit);
+  }
+  if (isList(field.type) && Array.isArray(value)) {
+    const child = field.type.children[0];
+    let out: unknown[] | undefined;
+    for (const [index, element] of value.entries()) {
+      const mapped = mapBlobField(
+        element,
+        child,
+        `${label}[${index}]`,
+        row,
+        visit,
+      );
+      if (mapped !== element) {
+        out ??= [...value];
+        out[index] = mapped;
+      }
+    }
+    return out ?? value;
+  }
+  return value;
+}
+
+/**
+ * Reads `Blob` / `File` values in the blob columns of `schema` into bytes so
+ * the synchronous conversion in {@link makeArrowTable} can accept them.
+ * Returns `data` itself when there is nothing to read.
+ */
+export async function resolveBlobInputs(
+  data: Array<Record<string, unknown>>,
+  schema?: SchemaLike,
+): Promise<Array<Record<string, unknown>>> {
+  if (schema === undefined || schema === null) {
+    return data;
+  }
+  const sanitized = sanitizeSchema(schema);
+  // Keyed by the Blob itself, so rows sharing one Blob (bare or as
+  // `{ data }`) read it once and share the bytes.
+  const reads = new Map<Blob, Promise<Uint8Array>>();
+  mapBlobInputs(data, sanitized, (value) => {
+    const source = blobToRead(value);
+    if (source !== undefined && !reads.has(source)) {
+      reads.set(
+        source,
+        source.arrayBuffer().then((buffer) => new Uint8Array(buffer)),
+      );
+    }
+    return value;
+  });
+  if (reads.size === 0) {
+    return data;
+  }
+  const bytes = new Map(
+    await Promise.all(
+      [...reads].map(async ([source, read]) => [source, await read] as const),
+    ),
+  );
+  return mapBlobInputs(data, sanitized, (value) => {
+    const source = blobToRead(value);
+    return source === undefined
+      ? value
+      : withBlobBytes(value, bytes.get(source) as Uint8Array);
+  });
+}
+
 function isObject(value: unknown): value is Record<string, unknown> {
   return (
     typeof value === "object" &&
@@ -1013,6 +1163,11 @@ export async function convertToTable(
     );
   }
 
+  processedData = await resolveBlobInputs(
+    processedData,
+    makeTableOptions?.schema,
+  );
+
   const table = makeArrowTable(processedData, makeTableOptions);
   return await applyEmbeddings(table, embeddings, makeTableOptions?.schema);
 }
@@ -1369,7 +1524,7 @@ export function ensureNestedFieldsExist(
     for (const field of schema.fields) {
       if (field.name in row) {
         if (
-          field.type.constructor.name === "Struct" &&
+          isPlainStructField(field) &&
           row[field.name] !== null &&
           row[field.name] !== undefined
         ) {
@@ -1386,10 +1541,9 @@ export function ensureNestedFieldsExist(
       } else {
         // Keep a missing struct valid while filling each of its children with
         // null. This is distinct from an explicitly null struct value.
-        completeRow[field.name] =
-          field.type.constructor.name === "Struct"
-            ? ensureStructFieldsExist({}, field.type as Struct)
-            : null;
+        completeRow[field.name] = isPlainStructField(field)
+          ? ensureStructFieldsExist({}, field.type as Struct)
+          : null;
       }
     }
 
@@ -1397,6 +1551,15 @@ export function ensureNestedFieldsExist(
   });
 }
 
+/**
+ * Blob fields are Arrow structs, but their values are bytes, URIs, or
+ * `{ data } | { uri }` inputs that the blob coercion in `makeArrowTable`
+ * handles, so they must not be filled in like ordinary structs.
+ */
+function isPlainStructField(field: Field): boolean {
+  return field.type.constructor.name === "Struct" && !isBlobField(field);
+}
+
 /**
  * Recursively ensures that all fields in a struct type exist in the data,
  * filling missing fields with null values.
@@ -1410,7 +1573,7 @@ function ensureStructFieldsExist(
   for (const childField of structType.children) {
     if (childField.name in data) {
       if (
-        childField.type.constructor.name === "Struct" &&
+        isPlainStructField(childField) &&
         data[childField.name] !== null &&
         data[childField.name] !== undefined
       ) {
@@ -1426,10 +1589,9 @@ function ensureStructFieldsExist(
     } else {
       // Keep a missing struct valid while filling each of its children with
       // null. This is distinct from an explicitly null struct value.
-      completeStruct[childField.name] =
-        childField.type.constructor.name === "Struct"
-          ? ensureStructFieldsExist({}, childField.type as Struct)
-          : null;
+      completeStruct[childField.name] = isPlainStructField(childField)
+        ? ensureStructFieldsExist({}, childField.type as Struct)
+        : null;
     }
   }
 

nodejs/lancedb/blob.ts

@@ -12,7 +12,46 @@ const DEDICATED_SIZE_THRESHOLD_KEY =
 const PACK_FILE_SIZE_THRESHOLD_KEY =
   "lance-encoding:blob-pack-file-size-threshold";
 
-export type BlobInput = {
+/**
+ * Bytes accepted for a blob value. `ArrayBuffer` is wrapped without copying.
+ * `Blob` and `File` are read with `arrayBuffer()`, which only the async write
+ * paths ({@link Connection.createTable}, {@link Table.add},
+ * {@link Table.mergeInsert}) can do.
+ */
+export type BlobData = Buffer | Uint8Array | ArrayBuffer | Blob;
+
+/** A URI accepted for a blob value. A `URL` is stored as its `href`. */
+export type BlobUri = string | URL;
+
+/**
+ * A value accepted for a `lance.blob.v2` column (see {@link blob}).
+ *
+ * Either inline bytes, a URI pointing at external bytes, or a struct that sets
+ * exactly one of `data` or `uri`. `null` and `undefined` write a null blob.
+ *
+ * @example
+ * ```ts
+ * import { pathToFileURL } from "node:url";
+ * import type { BlobInput } from "@lancedb/lancedb";
+ *
+ * const rows: { id: bigint; image: BlobInput }[] = [
+ *   { id: 1n, image: await (await fetch(url)).arrayBuffer() },
+ *   { id: 2n, image: pathToFileURL("/data/cat.png") },
+ *   { id: 3n, image: { data: new Blob(["hello"]) } },
+ * ];
+ * await table.add(rows);
+ * ```
+ */
+export type BlobInput =
+  | BlobData
+  | BlobUri
+  | { data: BlobData; uri?: null }
+  | { data?: null; uri: BlobUri }
+  | null
+  | undefined;
+
+/** A blob value normalized to the columns of the blob struct. */
+export type BlobValue = {
   data: Buffer | Uint8Array | null;
   uri: string | null;
 };
@@ -248,7 +287,77 @@ export type BlobRange = {
   end: bigint;
 };
 
-export function coerceBlobValue(value: unknown): BlobInput | null {
+/**
+ * Rewrites the synchronous-only widenings of {@link BlobInput} (`ArrayBuffer`,
+ * `URL`) into `Uint8Array` and URI strings, keeping the value's shape. Other
+ * values are returned unchanged and validated later by {@link coerceBlobValue}.
+ *
+ * Throws on `Blob` / `File`, which must be read into bytes first.
+ */
+export function normalizeBlobInput(value: unknown): unknown {
+  if (blobToRead(value) !== undefined) {
+    throw new Error(
+      "Blob and File values must be read asynchronously. Pass them to " +
+        "createTable, add, or mergeInsert, or convert them with " +
+        "`new Uint8Array(await value.arrayBuffer())`",
+    );
+  }
+  if (value instanceof ArrayBuffer) {
+    return new Uint8Array(value);
+  }
+  if (value instanceof URL) {
+    return value.href;
+  }
+  if (isBlobStruct(value)) {
+    // Read `data` and `uri` explicitly: spreading a class instance would drop
+    // fields defined as getters on its prototype.
+    const out: Record<string, unknown> = {
+      ...value,
+      data: value.data,
+      uri: value.uri,
+    };
+    let changed = false;
+    if (value.data instanceof ArrayBuffer) {
+      out.data = new Uint8Array(value.data);
+      changed = true;
+    }
+    if (value.uri instanceof URL) {
+      out.uri = value.uri.href;
+      changed = true;
+    }
+    return changed ? out : value;
+  }
+  return value;
+}
+
+/**
+ * Returns the `Blob` / `File` a blob value needs read, either the value itself
+ * or its `data` field, or `undefined` when there is nothing to read.
+ */
+export function blobToRead(value: unknown): Blob | undefined {
+  if (isBlobLike(value)) {
+    return value;
+  }
+  if (isBlobStruct(value) && isBlobLike(value.data)) {
+    return value.data;
+  }
+  return undefined;
+}
+
+/**
+ * Replaces the `Blob` that {@link blobToRead} found in `value` with `bytes`,
+ * keeping the value's shape.
+ */
+export function withBlobBytes(value: unknown, bytes: Uint8Array): unknown {
+  if (isBlobLike(value)) {
+    return bytes;
+  }
+  const record = value as Record<string, unknown>;
+  return { ...record, uri: record.uri, data: bytes };
+}
+
+export function coerceBlobValue(input: unknown): BlobValue | null {
+  const value = normalizeBlobInput(input);
   if (value == null) {
     return null;
   }
@@ -256,7 +365,9 @@ export function coerceBlobValue(value: unknown): BlobInput | null {
     return { data: value, uri: null };
   }
   if (ArrayBuffer.isView(value)) {
-    throw new Error("Blob data must be Buffer or Uint8Array");
+    throw new Error(
+      "Blob data must be Buffer, Uint8Array, ArrayBuffer, or Blob",
+    );
   }
   if (typeof value === "string") {
     if (value === "") {
@@ -264,23 +375,37 @@ export function coerceBlobValue(value: unknown): BlobInput | null {
     }
     return { data: null, uri: value };
   }
-  if (typeof value === "object") {
-    const record = value as Record<string, unknown>;
+  if (isBlobStruct(value)) {
+    const record = value;
     if (!("data" in record) && !("uri" in record)) {
       throw new Error(
         "Blob struct values must include a 'data' or 'uri' field",
       );
     }
+    const unknownKeys = Object.keys(record).filter(
+      (key) => key !== "data" && key !== "uri",
+    );
+    if (unknownKeys.length > 0) {
+      throw new Error(
+        `Blob struct values only support 'data' and 'uri', got ${unknownKeys
+          .map((key) => `'${key}'`)
+          .join(", ")}`,
+      );
+    }
     const uri = record.uri;
     if (uri === "") {
       throw new Error("Blob uri cannot be empty");
     }
     if (uri != null && typeof uri !== "string") {
-      throw new Error(`Blob uri must be a string or null, got ${typeof uri}`);
+      throw new Error(
+        `Blob uri must be a string, URL, or null, got ${typeof uri}`,
+      );
     }
     const data = record.data;
     if (data != null && !isBlobBytes(data)) {
-      throw new Error("Blob data must be Buffer, Uint8Array, or null");
+      throw new Error(
+        "Blob data must be Buffer, Uint8Array, ArrayBuffer, Blob, or null",
+      );
     }
     const bytes = (data as Buffer | Uint8Array | null | undefined) ?? null;
     const uriValue = uri ?? null;
@@ -292,7 +417,39 @@ export function coerceBlobValue(value: unknown): BlobInput | null {
     return { data: bytes, uri: uriValue };
   }
   throw new Error(
-    "Blob column values must be Buffer, Uint8Array, a URI string, null, or { data?, uri? }",
+    "Blob column values must be Buffer, Uint8Array, ArrayBuffer, Blob, a URI string or URL, null, or { data?, uri? }",
+  );
+}
+
+function isBlobLike(value: unknown): value is Blob {
+  return typeof Blob !== "undefined" && value instanceof Blob;
+}
+
+/**
+ * Whether `value` is a `{ data?, uri? }` struct: any plain object, or any
+ * other object with a `data` or `uri` field (such as a class instance), apart
+ * from the binary, `Blob`, and `URL` values that are blob inputs themselves.
+ */
+function isBlobStruct(value: unknown): value is Record<string, unknown> {
+  if (
+    typeof value !== "object" ||
+    value === null ||
+    Array.isArray(value) ||
+    ArrayBuffer.isView(value) ||
+    value instanceof ArrayBuffer ||
+    (typeof SharedArrayBuffer !== "undefined" &&
+      value instanceof SharedArrayBuffer) ||
+    isBlobLike(value) ||
+    value instanceof URL
+  ) {
+    return false;
+  }
+  const prototype = Object.getPrototypeOf(value);
+  return (
+    prototype === Object.prototype ||
+    prototype === null ||
+    "data" in value ||
+    "uri" in value
   );
 }
 

nodejs/lancedb/catalog.ts

@@ -5,9 +5,9 @@ import { Connection, LocalConnection } from "./connection";
 import { HeaderProvider } from "./header";
 import {
   JsHeaderProvider,
-  ListDatabasesResponse,
   Catalog as NativeCatalog,
   CatalogOptions as NativeCatalogOptions,
+  DatabaseNames as NativeDatabaseNames,
 } from "./native.js";
 import { OAuthConfig } from "./oauth";
 
@@ -21,7 +21,41 @@ export interface CatalogOptions
     | (() => Record<string, string> | Promise<Record<string, string>>);
 }
 
-export type { ListDatabasesResponse } from "./native.js";
+/**
+ * A lazy async iterator of database names with pagination state.
+ * Returned by {@link Catalog.listDatabases}.
+ */
+export class DatabaseNames implements AsyncIterableIterator<string> {
+  /** @hidden */
+  constructor(private readonly inner: NativeDatabaseNames) {}
+
+  [Symbol.asyncIterator](): AsyncIterableIterator<string> {
+    return this;
+  }
+
+  /** Fetch the next name, requesting another page only when needed. */
+  async next(): Promise<IteratorResult<string>> {
+    const value = await this.inner.next();
+    return value == null
+      ? { done: true, value: undefined }
+      : { done: false, value };
+  }
+
+  /** Number of names available without another REST request. */
+  numPageResults(): number {
+    return this.inner.numPageResults();
+  }
+
+  /**
+   * Token for the next REST request. Initially this is the supplied starting token.
+   * Returns undefined after the final page is fetched, even if names remain cached.
+   * Drain the cache before saving a token to avoid skipping names when resuming.
+   * A failed request terminates iteration and retains its token for a new iterator.
+   */
+  pageToken(): string | undefined {
+    return this.inner.pageToken() ?? undefined;
+  }
+}
 
 /** A remote catalog manages databases through the server's root namespace. */
 export class Catalog {
@@ -56,21 +90,35 @@ export class Catalog {
     await this.inner.dropDatabase(name, options.ignoreMissing);
   }
 
-  /** List one page of databases; pass pageToken from a response for the next page. */
-  async listDatabases(
-    options: { limit?: number; pageToken?: string } = {},
-  ): Promise<ListDatabasesResponse> {
+  /**
+   * Iterate lazily over all database names using `for await...of`.
+   * pageToken resumes from a saved token; omitted starts at the beginning.
+   * pageLimit limits each REST response, not the total; omitted uses the server default.
+   * Request errors are raised during iteration and terminate the iterator.
+   *
+   * @example
+   * ```ts
+   * for await (const name of catalog.listDatabases({ pageLimit: 20 })) {
+   *   console.log(name);
+   * }
+   * ```
+   */
+  listDatabases(
+    options: { pageLimit?: number; pageToken?: string } = {},
+  ): DatabaseNames {
     if (
-      options.limit !== undefined &&
-      (!Number.isInteger(options.limit) ||
-        options.limit <= 0 ||
-        options.limit > 2147483647)
+      options.pageLimit !== undefined &&
+      (!Number.isInteger(options.pageLimit) ||
+        options.pageLimit <= 0 ||
+        options.pageLimit > 2147483647)
     ) {
       throw new Error(
         "Database list limit must be an integer between 1 and 2147483647",
       );
     }
-    return this.inner.listDatabases(options.limit, options.pageToken);
+    return new DatabaseNames(
+      this.inner.listDatabases(options.pageToken, options.pageLimit),
+    );
   }
 }
 
@@ -82,7 +130,9 @@ export class Catalog {
  * ```ts
  * const catalog = await connectCatalog("https://my-server.example", { apiKey: "secret" });
  * const db = await catalog.createDatabase("analytics", { existOk: true });
- * const page = await catalog.listDatabases({ limit: 20 });
+ * for await (const name of catalog.listDatabases({ pageLimit: 20 })) {
+ *   console.log(name);
+ * }
  * ```
  */
 export async function connectCatalog(

nodejs/lancedb/connection.ts

@@ -8,6 +8,7 @@ import {
   fromTableToStreamBuffer,
   isArrowTable,
   makeArrowTable,
+  resolveBlobInputs,
 } from "./arrow";
 import {
   Table as ArrowTable,
@@ -324,10 +325,6 @@ export abstract class Connection {
    * The view is populated before creation returns. Set `withNoData` to create
    * only its definition and empty backing table. The view is a normal table:
    * it can be queried, indexed and searched, and it appears in `tableNames`.
-   * The source table must have stable row ids (create it with
-   * the `newTableEnableStableRowIds` storage option); they keep the view's
-   * provenance valid across source compactions and cannot be enabled after
-   * a table exists.
    */
   abstract createMaterializedView(
     name: string,
@@ -1133,7 +1130,11 @@ async function parseTableData(
   if (isArrowTable(data)) {
     table = sanitizeTable(data);
   } else {
-    table = makeArrowTable(data as Record<string, unknown>[], options);
+    const records = await resolveBlobInputs(
+      data as Record<string, unknown>[],
+      options?.schema,
+    );
+    table = makeArrowTable(records, options);
   }
   if (streaming) {
     const buf = await fromTableToStreamBuffer(

nodejs/lancedb/index.ts

@@ -84,7 +84,14 @@ export {
 } from "./arrow";
 
 export { blob, isBlobField, BlobFile } from "./blob";
-export type { BlobOptions, BlobRange, BlobReadOptions } from "./blob";
+export type {
+  BlobData,
+  BlobInput,
+  BlobOptions,
+  BlobRange,
+  BlobReadOptions,
+  BlobUri,
+} from "./blob";
 
 export {
   Connection,
@@ -639,6 +646,6 @@ export async function connectNamespace(
 export {
   Catalog,
   CatalogOptions,
-  ListDatabasesResponse,
+  DatabaseNames,
   connectCatalog,
 } from "./catalog";

nodejs/lancedb/materialized_view.ts

@@ -9,10 +9,9 @@ export const DEFINITION_META_KEY = "mv.definition";
 
 /**
  * The newest stored layout this version reads: `{"format": N, "query": "<SQL>"}`,
- * format 2 being a query with `GROUP BY`, format 3 native pairs and format 4
- * native dedup results.
+ * format 2 being a query with `GROUP BY`.
  */
-export const DEFINITION_FORMAT = 4;
+export const DEFINITION_FORMAT = 2;
 
 /**
  * The query that defines a materialized view, as stored:
@@ -188,23 +187,14 @@ export class MaterializedView {
   /**
    * Recompute the view from its source.
    *
-   * The refresh is incremental when the source's changes can be reconciled
-   * into the view -- rows added, changed or removed since the last one --
-   * and otherwise rebuilds. `full` forces a rebuild; `sourceVersion`
-   * refreshes to that source version instead of the latest.
-   *
-   * Concurrent refreshes of one view do not duplicate its rows. Two that
-   * plan the same source rows conflict on commit, and the loser throws
-   * rather than writing them a second time.
+   * Every refresh rebuilds the complete view. For local views,
+   * `sourceVersion` refreshes to that source version instead of the latest;
+   * remote SQL refreshes do not support source-version pinning.
    */
   async refresh(options?: {
-    full?: boolean;
     sourceVersion?: number;
   }): Promise<RefreshMaterializedViewResult> {
     validateNonNegativeInteger(options?.sourceVersion, "sourceVersion");
-    return await this.inner.refreshMaterializedView(
-      options?.full,
-      options?.sourceVersion,
-    );
+    return await this.inner.refreshMaterializedView(options?.sourceVersion);
   }
 }

nodejs/lancedb/schema.ts

@@ -69,12 +69,18 @@ class SchemaInferrer {
       this.addField(path, value, row);
     } else if (this.providedSchema === undefined) {
       this.updateInferredField(path, value, row, current);
+    } else if (current instanceof DeferredTypeEvidence) {
+      if (value != null) {
+        this.addSchemaField(this.providedSchema, path, value, row);
+      }
+    } else if (current instanceof FieldTree || DataType.isStruct(current)) {
+      assertStructValue(path, value, row);
     }
   }
 
   private addField(path: string[], value: unknown, row: number): void {
     if (this.providedSchema !== undefined) {
-      this.addSchemaField(this.providedSchema, path, row);
+      this.addSchemaField(this.providedSchema, path, value, row);
       return;
     }
 
@@ -95,15 +101,31 @@ class SchemaInferrer {
     }
   }
 
-  private addSchemaField(schema: Schema, path: string[], row: number): void {
+  private addSchemaField(
+    schema: Schema,
+    path: string[],
+    value: unknown,
+    row: number,
+  ): void {
     const field = fieldAtPath(schema, path);
     if (field === undefined) {
       throw new Error(
         `Found field not in schema: ${path.join(".")} at row ${row}`,
       );
     }
+    if (DataType.isStruct(field.type)) {
+      assertStructValue(path, value, row);
+    }
 
-    const conflict = this.fields.set(path, field.type);
+    // Nulls are kept as deferred evidence so that a struct field seen first as
+    // null can be replaced by a record in a later row, just as a record seen
+    // first already ignores later nulls.
+    const conflict = this.fields.set(
+      path,
+      DeferredTypeEvidence.from(value, row) ?? field.type,
+      (existing) =>
+        existing instanceof DeferredTypeEvidence && existing.isOnlyNulls(),
+    );
     if (conflict !== undefined) {
       throw branchConflictError(conflict, row, "Struct");
     }
@@ -483,6 +505,26 @@ function assertSupportedValue(
   }
 }
 
+/**
+ * Rejects primitives, arrays, and typed arrays given for a struct field, which
+ * would otherwise be written as a struct with every child null.
+ */
+function assertStructValue(path: string[], value: unknown, row: number): void {
+  if (
+    value == null ||
+    (typeof value === "object" &&
+      !Array.isArray(value) &&
+      !ArrayBuffer.isView(value))
+  ) {
+    return;
+  }
+  const kind =
+    typeof value === "object" ? value.constructor.name : typeof value;
+  throw new Error(
+    `Expected a struct value for field ${path.join(".")} at row ${row}, got ${kind}.`,
+  );
+}
+
 function fieldAtPath(schema: Schema, path: string[]): Field | undefined {
   let fields = schema.fields;
   let field: Field | undefined;

nodejs/lancedb/table.ts

@@ -225,9 +225,8 @@ export interface LsmWriteSpec {
   /** Bucket variant: the number of buckets, in `[1, 1024]`. */
   numBuckets?: number;
   /**
-   * Indexes the MemWAL keeps up to date. Omit to maintain every supported
-   * index, resolved on install — a snapshot, so indexes created later are not
-   * maintained. Pass `[]` for none.
+   * Indexes the MemWAL keeps up to date. Omit for every index the table has,
+   * including ones created later. Pass `[]` for none.
    */
   maintainedIndexes?: string[];
   /** Default `ShardWriter` configuration recorded in the MemWAL index. */
@@ -685,7 +684,6 @@ export abstract class Table {
    * @ignore
    */
   abstract refreshMaterializedView(
-    full?: boolean,
     sourceVersion?: number,
   ): Promise<RefreshMaterializedViewResult>;
 
@@ -767,8 +765,8 @@ export abstract class Table {
    * ({@link Table#setUnenforcedPrimaryKey}); bucket sharding additionally
    * requires it to be the single column being bucketed.
    *
-   * Omitting `maintainedIndexes` maintains every index on the table, resolved
-   * here, failing if one cannot be maintained — name them to install anyway.
+   * Omitting `maintainedIndexes` maintains every index the table has,
+   * including ones created later, and skips a kind the MemWAL cannot maintain.
    * Naming them pins an exact set, and a still-building index is rejected
    * rather than quietly omitted.
    * @param {LsmWriteSpec} spec The sharding spec to install.
@@ -798,10 +796,9 @@ export abstract class Table {
    *
    * Resolves to `undefined` when the MemWAL LSM write path is not enabled (no
    * spec has been set, or it was removed with {@link Table#unsetLsmWriteSpec}).
-   * The returned spec mirrors what was passed to
-   * {@link Table#setLsmWriteSpec}, except that `maintainedIndexes` always
-   * reports the concrete list resolved when the spec was set — `undefined`
-   * never round-trips.
+   * 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<LsmWriteSpec | undefined>}
    */
   abstract getLsmWriteSpec(): Promise<LsmWriteSpec | undefined>;
@@ -1420,10 +1417,9 @@ export class LocalTable extends Table {
   }
 
   async refreshMaterializedView(
-    full?: boolean,
     sourceVersion?: number,
   ): Promise<RefreshMaterializedViewResult> {
-    return await this.inner.refreshMaterializedView(full, sourceVersion);
+    return await this.inner.refreshMaterializedView(sourceVersion);
   }
 
   async materializedViewDefinition(): Promise<string> {

nodejs/npm/darwin-arm64/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-darwin-arm64",
-	"version": "0.40.0-beta.12",
+	"version": "0.40.0-beta.13",
 	"os": ["darwin"],
 	"cpu": ["arm64"],
 	"main": "lancedb.darwin-arm64.node",

nodejs/npm/linux-arm64-gnu/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-linux-arm64-gnu",
-	"version": "0.40.0-beta.12",
+	"version": "0.40.0-beta.13",
 	"os": ["linux"],
 	"cpu": ["arm64"],
 	"main": "lancedb.linux-arm64-gnu.node",

nodejs/npm/linux-arm64-musl/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-linux-arm64-musl",
-	"version": "0.40.0-beta.12",
+	"version": "0.40.0-beta.13",
 	"os": ["linux"],
 	"cpu": ["arm64"],
 	"main": "lancedb.linux-arm64-musl.node",

nodejs/npm/linux-x64-gnu/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-linux-x64-gnu",
-	"version": "0.40.0-beta.12",
+	"version": "0.40.0-beta.13",
 	"os": ["linux"],
 	"cpu": ["x64"],
 	"main": "lancedb.linux-x64-gnu.node",

nodejs/npm/linux-x64-musl/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-linux-x64-musl",
-	"version": "0.40.0-beta.12",
+	"version": "0.40.0-beta.13",
 	"os": ["linux"],
 	"cpu": ["x64"],
 	"main": "lancedb.linux-x64-musl.node",

nodejs/npm/win32-arm64-msvc/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@lancedb/lancedb-win32-arm64-msvc",
-  "version": "0.40.0-beta.12",
+  "version": "0.40.0-beta.13",
   "os": [
     "win32"
   ],

nodejs/npm/win32-x64-msvc/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-win32-x64-msvc",
-	"version": "0.40.0-beta.12",
+	"version": "0.40.0-beta.13",
 	"os": ["win32"],
 	"cpu": ["x64"],
 	"main": "lancedb.win32-x64-msvc.node",

nodejs/package.json

@@ -11,7 +11,7 @@
     "ann"
   ],
   "private": false,
-  "version": "0.40.0-beta.12",
+  "version": "0.40.0-beta.13",
   "main": "dist/index.js",
   "exports": {
     ".": "./dist/index.js",

nodejs/src/catalog.rs

@@ -1,12 +1,12 @@
 // SPDX-License-Identifier: Apache-2.0
 // SPDX-FileCopyrightText: Copyright The LanceDB Authors
 
-use std::sync::Arc;
+use std::sync::{Arc, Mutex};
 use std::time::Duration;
 
-use lancedb::catalog::{
-    CatalogConnection, CreateDatabaseRequest, DropDatabaseRequest, ListDatabasesRequest,
-};
+use futures::{StreamExt, future::poll_fn};
+
+use lancedb::catalog::{CatalogConnection, CreateDatabaseRequest, DropDatabaseRequest};
 use napi::bindgen_prelude::*;
 use napi_derive::napi;
 
@@ -25,10 +25,37 @@ pub struct CatalogOptions {
     pub oauth_config: Option<OAuthConfig>,
 }
 
-#[napi(object)]
-pub struct ListDatabasesResponse {
-    pub databases: Vec<String>,
-    pub page_token: Option<String>,
+/// A lazy iterator over database names with inspectable pagination state.
+#[napi]
+pub struct DatabaseNames {
+    inner: Mutex<lancedb::catalog::DatabaseNames>,
+    next_lock: futures::lock::Mutex<()>,
+}
+
+#[napi]
+impl DatabaseNames {
+    /// Number of names cached without another REST request.
+    #[napi]
+    pub fn num_page_results(&self) -> u32 {
+        self.inner.lock().unwrap().num_page_results() as u32
+    }
+
+    /// Continuation token for the next REST request.
+    #[napi]
+    pub fn page_token(&self) -> Option<String> {
+        self.inner.lock().unwrap().page_token().map(str::to_owned)
+    }
+
+    /// Fetch the next name, returning None when exhausted.
+    #[napi]
+    pub async fn next(&self) -> Result<Option<String>> {
+        // Serialize advances but allow synchronous inspection between polls.
+        let _guard = self.next_lock.lock().await;
+        poll_fn(|cx| self.inner.lock().unwrap().poll_next_unpin(cx))
+            .await
+            .transpose()
+            .default_error()
+    }
 }
 
 #[napi]
@@ -106,18 +133,14 @@ impl Catalog {
             .default_error()
     }
     #[napi]
-    pub async fn list_databases(
+    pub fn list_databases(
         &self,
-        limit: Option<u32>,
         page_token: Option<String>,
-    ) -> Result<ListDatabasesResponse> {
-        let mut request = ListDatabasesRequest::default();
-        request.limit = limit;
-        request.page_token = page_token;
-        let response = self.inner.list_databases(request).await.default_error()?;
-        Ok(ListDatabasesResponse {
-            databases: response.databases,
-            page_token: response.page_token,
-        })
+        page_limit: Option<u32>,
+    ) -> DatabaseNames {
+        DatabaseNames {
+            inner: Mutex::new(self.inner.list_databases(page_token, page_limit)),
+            next_lock: futures::lock::Mutex::new(()),
+        }
     }
 }

nodejs/src/table.rs

@@ -469,13 +469,12 @@ impl Table {
     #[napi(catch_unwind)]
     pub async fn refresh_materialized_view(
         &self,
-        full: Option<bool>,
         source_version: Option<i64>,
     ) -> napi::Result<RefreshMaterializedViewResult> {
         let view = lancedb::MaterializedView::from_table(self.inner_ref()?.clone())
             .await
             .default_error()?;
-        let mut builder = view.refresh().full(full.unwrap_or(false));
+        let mut builder = view.refresh();
         if let Some(version) = source_version {
             let version = u64::try_from(version).map_err(|_| {
                 napi::Error::from_reason("sourceVersion must be a non-negative integer")
@@ -492,11 +491,9 @@ impl Table {
         let view = lancedb::MaterializedView::from_table(inner)
             .await
             .default_error()?;
-        view.definition().to_json().map_err(|err| {
-            napi::Error::from_reason(format!(
-                "failed to serialize materialized-view definition: {err}"
-            ))
-        })
+        Ok(lancedb::materialized_view::definition_metadata_from_sql(
+            view.definition_sql(),
+        ))
     }
 
     #[napi(catch_unwind)]
@@ -1614,7 +1611,7 @@ impl From<lancedb::function::FunctionErrors> for FunctionErrors {
 
 #[napi(object)]
 pub struct RefreshMaterializedViewResult {
-    /// How the view was brought up to date: "rebuild", "incremental" or "no_op".
+    /// How the view was brought up to date: "rebuild".
     pub mode: String,
     pub rows_written: i64,
     pub source_version: i64,
@@ -1625,8 +1622,6 @@ impl From<lancedb::RefreshMaterializedViewResult> for RefreshMaterializedViewRes
     fn from(value: lancedb::RefreshMaterializedViewResult) -> Self {
         let mode = match value.mode {
             lancedb::RefreshMode::Rebuild => "rebuild",
-            lancedb::RefreshMode::Incremental => "incremental",
-            lancedb::RefreshMode::NoOp => "no_op",
         };
         Self {
             mode: mode.to_string(),