diff --git a/nodejs/__test__/arrow.test.ts b/nodejs/__test__/arrow.test.ts index 6c1d5ecba..df02b4936 100644 --- a/nodejs/__test__/arrow.test.ts +++ b/nodejs/__test__/arrow.test.ts @@ -601,6 +601,103 @@ describe.each([arrow15, arrow16, arrow17, arrow18])( ); }); + it.each([ + ["ArrayBuffer", new Uint8Array([104]).buffer], + ["URL", new URL("https://example.com/")], + ["empty object", {}], + [ + "class instance with enumerable fields", + Object.assign(new Blob(["x"]), { position: 0 }), + ], + ])("rejects %s values without a schema", (_name, value) => { + expect(() => makeArrowTable([{ id: 1, value }])).toThrow( + /field value at row 0/, + ); + expect(() => makeArrowTable([{ value }])).toThrow( + /field value at row 0/, + ); + }); + + it.each([ + ["Binary", new Binary(), new Uint8Array([104]).buffer], + ["Utf8", new Utf8(), new URL("https://example.com/")], + ])("rejects %s values with a schema", (_name, type, value) => { + const schema = new Schema([new Field("value", type, true)]); + expect(() => makeArrowTable([{ value }], { schema })).toThrow( + /field value at row 0/, + ); + }); + + it("rejects unsupported values in nested struct fields", () => { + const schema = new Schema([ + new Field( + "metadata", + new Struct([new Field("bytes", new Binary(), true)]), + true, + ), + ]); + expect(() => + makeArrowTable( + [{ metadata: { bytes: new Uint8Array([104]).buffer } }], + { schema }, + ), + ).toThrow(/field metadata\.bytes at row 0/); + }); + + it("rejects unsupported objects in list elements", () => { + const schema = new Schema([ + new Field( + "images", + new List(new Field("item", new Binary(), true)), + true, + ), + ]); + expect(() => + makeArrowTable([{ images: [new Blob(["abc"])] }], { schema }), + ).toThrow(/field images\[0\] at row 0/); + }); + + it("rejects unsupported objects in nested lists", () => { + const schema = new Schema([ + new Field( + "images", + new List( + new Field( + "item", + new List(new Field("item", new Binary(), true)), + true, + ), + ), + true, + ), + ]); + expect(() => + makeArrowTable([{ images: [[new Blob(["abc"])]] }], { schema }), + ).toThrow(/field images\[0\]\[0\] at row 0/); + }); + + it("rejects unsupported nested struct fields in lists", () => { + const schema = new Schema([ + new Field( + "items", + new List( + new Field( + "item", + new Struct([new Field("bytes", new Binary(), true)]), + true, + ), + ), + true, + ), + ]); + expect(() => + makeArrowTable( + [{ items: [{ bytes: new Uint8Array([104]).buffer }] }], + { schema }, + ), + ).toThrow(/field items\[0\]\.bytes at row 0/); + }); + it("will ignore generated dictionary IDs when comparing inferred types", function () { const table = makeArrowTable([{ str: "a" }, { str: "b" }], { dictionaryEncodeStrings: true, diff --git a/nodejs/__test__/blob.test.ts b/nodejs/__test__/blob.test.ts index e8e9357e3..5c01897b2 100644 --- a/nodejs/__test__/blob.test.ts +++ b/nodejs/__test__/blob.test.ts @@ -102,6 +102,27 @@ 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") }], + ])("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", () => { + 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.", + ); + }); + it("coerces Buffer input onto a blob field", () => { const schema = new Schema([ new Field("id", new Int64(), true), diff --git a/nodejs/__test__/table.test.ts b/nodejs/__test__/table.test.ts index 4142af696..ac8904c16 100644 --- a/nodejs/__test__/table.test.ts +++ b/nodejs/__test__/table.test.ts @@ -2412,6 +2412,23 @@ describe("when dealing with blob columns", () => { tmpDir.removeCallback(); }); + it("rejects unsupported blob values in createTable and add", async () => { + const db = await connect(tmpDir.name); + const schema = new Schema([ + new Field("id", new Int64(), true), + blob("image"), + ]); + const row = { id: 1n, image: new Uint8Array([104]).buffer }; + + await expect(db.createTable("invalid", [row], { schema })).rejects.toThrow( + /field image at row 0/, + ); + + const table = await db.createEmptyTable("empty", schema); + await expect(table.add([row])).rejects.toThrow(/field image at row 0/); + await expect(table.countRows()).resolves.toBe(0); + }); + it("discovers blob columns", async () => { const { table } = await openBlobTable(); expect(await table.blobColumns()).toEqual(["image"]); diff --git a/nodejs/lancedb/schema.ts b/nodejs/lancedb/schema.ts index 3a3ee9316..724a81fc9 100644 --- a/nodejs/lancedb/schema.ts +++ b/nodejs/lancedb/schema.ts @@ -62,6 +62,8 @@ class SchemaInferrer { } private observe(path: string[], value: unknown, row: number): void { + assertSupportedValue(value, path.join("."), row); + const current = this.fields.get(path); if (current === undefined) { this.addField(path, value, row); @@ -430,19 +432,57 @@ function* recordPathsAndValues( } function isRecord(value: unknown): value is Record { + if (typeof value !== "object" || value === null) { + return false; + } + const prototype = Object.getPrototypeOf(value); return ( - typeof value === "object" && - value !== null && - !Array.isArray(value) && - !(value instanceof RegExp) && - !(value instanceof Date) && - !(value instanceof Set) && - !(value instanceof Map) && - !(value instanceof Buffer) && - !ArrayBuffer.isView(value) + (prototype === Object.prototype || prototype === null) && + Object.keys(value).length > 0 ); } +function isUnsupportedObject(value: unknown): boolean { + if ( + typeof value !== "object" || + value === null || + Array.isArray(value) || + isRecord(value) + ) { + return false; + } + // Preserve the object types already passed through to Arrow. Other objects + // must not disappear while walking struct fields. + return !( + value instanceof RegExp || + value instanceof Date || + value instanceof Set || + value instanceof Map || + Buffer.isBuffer(value) || + ArrayBuffer.isView(value) + ); +} + +function assertSupportedValue( + value: unknown, + field: string, + row: number, +): void { + if (Array.isArray(value)) { + for (const [index, item] of value.entries()) { + assertSupportedValue(item, `${field}[${index}]`, row); + } + } else if (isRecord(value)) { + for (const [name, child] of Object.entries(value)) { + assertSupportedValue(child, `${field}.${name}`, row); + } + } else if (isUnsupportedObject(value)) { + throw new Error( + `Unsupported object value for field ${field} at row ${row}.`, + ); + } +} + function fieldAtPath(schema: Schema, path: string[]): Field | undefined { let fields = schema.fields; let field: Field | undefined;