|
| 1 | +// Licensed to the Apache Software Foundation (ASF) under one |
| 2 | +// or more contributor license agreements. See the NOTICE file |
| 3 | +// distributed with this work for additional information |
| 4 | +// regarding copyright ownership. The ASF licenses this file |
| 5 | +// to you under the Apache License, Version 2.0 (the |
| 6 | +// "License"); you may not use this file except in compliance |
| 7 | +// with the License. You may obtain a copy of the License at |
| 8 | +// |
| 9 | +// http://www.apache.org/licenses/LICENSE-2.0 |
| 10 | +// |
| 11 | +// Unless required by applicable law or agreed to in writing, |
| 12 | +// software distributed under the License is distributed on an |
| 13 | +// "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY |
| 14 | +// KIND, either express or implied. See the License for the |
| 15 | +// specific language governing permissions and limitations |
| 16 | +// under the License. |
| 17 | + |
| 18 | +import '../../../jest-extensions.js'; |
| 19 | +import { |
| 20 | + Field, |
| 21 | + Float64, |
| 22 | + Int32, |
| 23 | + makeData, |
| 24 | + RecordBatch, |
| 25 | + RecordBatchStreamWriter, |
| 26 | + Schema, |
| 27 | + Struct, |
| 28 | + Utf8, |
| 29 | +} from 'apache-arrow'; |
| 30 | + |
| 31 | +function createBatch(schema: Schema, data: Record<string, any>, length: number): RecordBatch { |
| 32 | + const children = schema.fields.map(field => { |
| 33 | + const d = data[field.name]; |
| 34 | + if (field.type instanceof Utf8) { |
| 35 | + return makeData({ |
| 36 | + type: field.type, |
| 37 | + data: Buffer.from(d.values), |
| 38 | + valueOffsets: new Int32Array(d.offsets), |
| 39 | + }); |
| 40 | + } |
| 41 | + return makeData({ type: field.type, data: d }); |
| 42 | + }); |
| 43 | + const structData = makeData({ |
| 44 | + type: new Struct(schema.fields), |
| 45 | + length, |
| 46 | + nullCount: 0, |
| 47 | + children, |
| 48 | + }); |
| 49 | + return new RecordBatch(schema, structData); |
| 50 | +} |
| 51 | + |
| 52 | +describe('RecordBatchStreamWriter schema mismatch behavior', () => { |
| 53 | + |
| 54 | + const schemaA = new Schema([ |
| 55 | + new Field('id', new Int32()), |
| 56 | + new Field('name', new Utf8()), |
| 57 | + ]); |
| 58 | + |
| 59 | + const schemaB = new Schema([ |
| 60 | + new Field('x', new Float64()), |
| 61 | + new Field('y', new Float64()), |
| 62 | + ]); |
| 63 | + |
| 64 | + function makeBatchA(): RecordBatch { |
| 65 | + return createBatch(schemaA, { |
| 66 | + id: new Int32Array([1, 2, 3]), |
| 67 | + name: { values: 'foobarbaz', offsets: [0, 3, 6, 9] }, |
| 68 | + }, 3); |
| 69 | + } |
| 70 | + |
| 71 | + function makeBatchB(): RecordBatch { |
| 72 | + return createBatch(schemaB, { |
| 73 | + x: new Float64Array([1.1, 2.2]), |
| 74 | + y: new Float64Array([3.3, 4.4]), |
| 75 | + }, 2); |
| 76 | + } |
| 77 | + |
| 78 | + test('default (autoDestroy=true): writing a mismatched batch throws an error', () => { |
| 79 | + const writer = new RecordBatchStreamWriter(); // autoDestroy defaults to true |
| 80 | + const batchA = makeBatchA(); |
| 81 | + const batchB = makeBatchB(); |
| 82 | + |
| 83 | + // Write the first batch — this establishes the schema |
| 84 | + writer.write(batchA); |
| 85 | + |
| 86 | + // Write a batch with a different schema — this should throw |
| 87 | + expect(() => writer.write(batchB)).toThrow( |
| 88 | + 'RecordBatch schema does not match the writer\'s schema.' |
| 89 | + ); |
| 90 | + }); |
| 91 | + |
| 92 | + test('autoDestroy=false: writing a mismatched batch resets the schema and writes both batches as separate streams', async () => { |
| 93 | + const writer = new RecordBatchStreamWriter({ autoDestroy: false }); |
| 94 | + const batchA = makeBatchA(); |
| 95 | + const batchB = makeBatchB(); |
| 96 | + |
| 97 | + // Write first batch |
| 98 | + writer.write(batchA); |
| 99 | + |
| 100 | + // Write batch with different schema — with autoDestroy=false, |
| 101 | + // this resets the writer to the new schema and writes the batch |
| 102 | + writer.write(batchB); |
| 103 | + writer.close(); |
| 104 | + |
| 105 | + // The output contains two separate IPC streams. |
| 106 | + // RecordBatchReader.readAll should be able to read both. |
| 107 | + const { RecordBatchReader } = await import('apache-arrow'); |
| 108 | + const buffer = await writer.toUint8Array(); |
| 109 | + const readers = []; |
| 110 | + for await (const reader of RecordBatchReader.readAll(buffer)) { |
| 111 | + readers.push(new (await import('apache-arrow')).Table(await reader.readAll())); |
| 112 | + } |
| 113 | + |
| 114 | + expect(readers).toHaveLength(2); |
| 115 | + expect(readers[0].numRows).toBe(3); |
| 116 | + expect(readers[0].schema.fields.map(f => f.name)).toEqual(['id', 'name']); |
| 117 | + expect(readers[1].numRows).toBe(2); |
| 118 | + expect(readers[1].schema.fields.map(f => f.name)).toEqual(['x', 'y']); |
| 119 | + }); |
| 120 | + |
| 121 | + test('writing to a closed writer throws an error', () => { |
| 122 | + const writer = new RecordBatchStreamWriter(); |
| 123 | + const batchA = makeBatchA(); |
| 124 | + |
| 125 | + writer.write(batchA); |
| 126 | + writer.close(); |
| 127 | + |
| 128 | + // Now try to write again — the writer is closed |
| 129 | + expect(() => writer.write(makeBatchA())).toThrow(); |
| 130 | + }); |
| 131 | + |
| 132 | + test('schema mismatch error includes both schemas', () => { |
| 133 | + const writer = new RecordBatchStreamWriter(); |
| 134 | + const batchA = makeBatchA(); |
| 135 | + const batchB = makeBatchB(); |
| 136 | + |
| 137 | + writer.write(batchA); |
| 138 | + |
| 139 | + expect(() => writer.write(batchB)).toThrow(/Expected:.*\n.*Received:/); |
| 140 | + }); |
| 141 | +}); |
0 commit comments