|
| 1 | +/* |
| 2 | + * Copyright (2025) The Delta Lake Project Authors. |
| 3 | + * |
| 4 | + * Licensed under the Apache License, Version 2.0 (the "License"); |
| 5 | + * you may not use this file except in compliance with the License. |
| 6 | + * You may obtain a copy of the License at |
| 7 | + * |
| 8 | + * http://www.apache.org/licenses/LICENSE-2.0 |
| 9 | + * |
| 10 | + * Unless required by applicable law or agreed to in writing, software |
| 11 | + * distributed under the License is distributed on an "AS IS" BASIS, |
| 12 | + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. |
| 13 | + * See the License for the specific language governing permissions and |
| 14 | + * limitations under the License. |
| 15 | + */ |
| 16 | +package io.delta.kernel.internal |
| 17 | + |
| 18 | +import java.util |
| 19 | +import java.util.{HashMap, Map} |
| 20 | +import java.util.Optional |
| 21 | + |
| 22 | +import scala.collection.JavaConverters._ |
| 23 | + |
| 24 | +import io.delta.kernel.data.Row |
| 25 | +import io.delta.kernel.internal.annotation.VisibleForTesting |
| 26 | +import io.delta.kernel.internal.data.GenericRow |
| 27 | +import io.delta.kernel.internal.replay.PageToken |
| 28 | +import io.delta.kernel.test.MockFileSystemClientUtils |
| 29 | +import io.delta.kernel.types._ |
| 30 | + |
| 31 | +import org.scalatest.funsuite.AnyFunSuite |
| 32 | + |
| 33 | +class PageTokenSuite extends AnyFunSuite with MockFileSystemClientUtils { |
| 34 | + |
| 35 | + private val TEST_FILE_NAME = "test_file.json" |
| 36 | + private val TEST_ROW_INDEX = 42L |
| 37 | + private val TEST_SIDECAR_INDEX = Optional.of(java.lang.Long.valueOf(5L)) |
| 38 | + private val TEST_KERNEL_VERSION = "4.0.0" |
| 39 | + private val TEST_TABLE_PATH = "/path/to/table" |
| 40 | + private val TEST_TABLE_VERSION = 5L |
| 41 | + private val TEST_PREDICATE_HASH = 123L |
| 42 | + private val TEST_LOG_SEGMENT_HASH = 456L |
| 43 | + |
| 44 | + private val expectedPageToken = new PageToken( |
| 45 | + TEST_FILE_NAME, |
| 46 | + TEST_ROW_INDEX, |
| 47 | + TEST_SIDECAR_INDEX, |
| 48 | + TEST_KERNEL_VERSION, |
| 49 | + TEST_TABLE_PATH, |
| 50 | + TEST_TABLE_VERSION, |
| 51 | + TEST_PREDICATE_HASH, |
| 52 | + TEST_LOG_SEGMENT_HASH) |
| 53 | + |
| 54 | + private val rowData: Map[Integer, Object] = new HashMap() |
| 55 | + rowData.put(0, TEST_FILE_NAME) |
| 56 | + rowData.put(1, TEST_ROW_INDEX.asInstanceOf[Object]) |
| 57 | + rowData.put(2, TEST_SIDECAR_INDEX.get()) |
| 58 | + rowData.put(3, TEST_KERNEL_VERSION) |
| 59 | + rowData.put(4, TEST_TABLE_PATH) |
| 60 | + rowData.put(5, TEST_TABLE_VERSION.asInstanceOf[Object]) |
| 61 | + rowData.put(6, TEST_PREDICATE_HASH.asInstanceOf[Object]) |
| 62 | + rowData.put(7, TEST_LOG_SEGMENT_HASH.asInstanceOf[Object]) |
| 63 | + |
| 64 | + val expectedRow = new GenericRow(PageToken.PAGE_TOKEN_SCHEMA, rowData) |
| 65 | + |
| 66 | + test("PageToken.fromRow with valid data") { |
| 67 | + val pageToken = PageToken.fromRow(expectedRow) |
| 68 | + assert(pageToken.equals(expectedPageToken)) |
| 69 | + } |
| 70 | + |
| 71 | + test("PageToken.toRow with valid data") { |
| 72 | + val row = expectedPageToken.toRow |
| 73 | + assert(row.getSchema.equals(PageToken.PAGE_TOKEN_SCHEMA)) |
| 74 | + |
| 75 | + assert(row.getString(0) == TEST_FILE_NAME) |
| 76 | + assert(row.getLong(1) == TEST_ROW_INDEX) |
| 77 | + assert(Optional.of(row.getLong(2)) == TEST_SIDECAR_INDEX) |
| 78 | + assert(row.getString(3) == TEST_KERNEL_VERSION) |
| 79 | + assert(row.getString(4) == TEST_TABLE_PATH) |
| 80 | + assert(row.getLong(5) == TEST_TABLE_VERSION) |
| 81 | + assert(row.getLong(6) == TEST_PREDICATE_HASH) |
| 82 | + assert(row.getLong(7) == TEST_LOG_SEGMENT_HASH) |
| 83 | + } |
| 84 | + |
| 85 | + test("E2E: PageToken round-trip: toRow -> fromRow") { |
| 86 | + val row = expectedPageToken.toRow |
| 87 | + val reconstructedPageToken = PageToken.fromRow(row) |
| 88 | + assert(reconstructedPageToken.equals(expectedPageToken)) |
| 89 | + } |
| 90 | + |
| 91 | + test("PageToken.fromRow throws exception when input row schema has invalid field name") { |
| 92 | + val invalidSchema = new StructType() |
| 93 | + .add("wrongFieldName", StringType.STRING) |
| 94 | + .add("lastReturnedRowIndex", LongType.LONG) |
| 95 | + .add("lastReadSidecarFileIdx", LongType.LONG) |
| 96 | + .add("kernelVersion", StringType.STRING) |
| 97 | + .add("tablePath", StringType.STRING) |
| 98 | + .add("tableVersion", LongType.LONG) |
| 99 | + .add("predicateHash", LongType.LONG) |
| 100 | + .add("logSegmentHash", LongType.LONG) |
| 101 | + |
| 102 | + val invalidRowData: Map[Integer, Object] = new HashMap() |
| 103 | + invalidRowData.put(0, TEST_FILE_NAME) |
| 104 | + invalidRowData.put(1, TEST_ROW_INDEX.asInstanceOf[Object]) |
| 105 | + invalidRowData.put(2, TEST_SIDECAR_INDEX) |
| 106 | + invalidRowData.put(3, TEST_KERNEL_VERSION) |
| 107 | + invalidRowData.put(4, TEST_TABLE_PATH) |
| 108 | + invalidRowData.put(5, TEST_TABLE_VERSION.asInstanceOf[Object]) |
| 109 | + invalidRowData.put(6, TEST_PREDICATE_HASH.asInstanceOf[Object]) |
| 110 | + invalidRowData.put(7, TEST_LOG_SEGMENT_HASH.asInstanceOf[Object]) |
| 111 | + |
| 112 | + val row = new GenericRow(invalidSchema, invalidRowData) |
| 113 | + val exception = intercept[IllegalArgumentException] { |
| 114 | + PageToken.fromRow(row) |
| 115 | + } |
| 116 | + assert(exception.getMessage.contains( |
| 117 | + "Invalid Page Token: input row schema does not match expected PageToken schema")) |
| 118 | + } |
| 119 | + |
| 120 | + test("PageToken.fromRow throws exception when input row schema has wrong data type") { |
| 121 | + val invalidSchema = new StructType() |
| 122 | + .add("lastReadLogFileName", StringType.STRING) |
| 123 | + .add("lastReturnedRowIndex", LongType.LONG) |
| 124 | + .add("lastReadSidecarFileIdx", StringType.STRING) // should be long type |
| 125 | + .add("kernelVersion", StringType.STRING) |
| 126 | + .add("tablePath", StringType.STRING) |
| 127 | + .add("tableVersion", LongType.LONG) |
| 128 | + .add("predicateHash", LongType.LONG) |
| 129 | + .add("logSegmentHash", LongType.LONG) |
| 130 | + |
| 131 | + val invalidRowData: Map[Integer, Object] = new HashMap() |
| 132 | + invalidRowData.put(0, TEST_FILE_NAME) |
| 133 | + invalidRowData.put(1, TEST_ROW_INDEX.asInstanceOf[Object]) |
| 134 | + invalidRowData.put(2, TEST_SIDECAR_INDEX) |
| 135 | + invalidRowData.put(3, TEST_KERNEL_VERSION) |
| 136 | + invalidRowData.put(4, TEST_TABLE_PATH) |
| 137 | + invalidRowData.put(5, TEST_TABLE_VERSION.asInstanceOf[Object]) |
| 138 | + invalidRowData.put(6, TEST_PREDICATE_HASH.asInstanceOf[Object]) |
| 139 | + invalidRowData.put(7, TEST_LOG_SEGMENT_HASH.asInstanceOf[Object]) |
| 140 | + |
| 141 | + val row = new GenericRow(invalidSchema, invalidRowData) |
| 142 | + val exception = intercept[IllegalArgumentException] { |
| 143 | + PageToken.fromRow(row) |
| 144 | + } |
| 145 | + assert(exception.getMessage.contains( |
| 146 | + "Invalid Page Token: input row schema does not match expected PageToken schema")) |
| 147 | + } |
| 148 | + |
| 149 | + test("PageToken.fromRow accepts the case sidecar field is null") { |
| 150 | + val nullSidecarData: Map[Integer, Object] = new HashMap(rowData) |
| 151 | + nullSidecarData.put(2, null) |
| 152 | + val nullSidecarRow = new GenericRow(PageToken.PAGE_TOKEN_SCHEMA, nullSidecarData) |
| 153 | + val pageToken = PageToken.fromRow(nullSidecarRow) |
| 154 | + assert(pageToken.getLastReadSidecarFileIdx == Optional.empty()) |
| 155 | + } |
| 156 | + |
| 157 | + test("PageToken.fromRow throws exception when required field is null") { |
| 158 | + val invalidData: Map[Integer, Object] = new HashMap(rowData) |
| 159 | + invalidData.put(3, null) |
| 160 | + val invalidRow = new GenericRow(PageToken.PAGE_TOKEN_SCHEMA, invalidData) |
| 161 | + val exception = intercept[IllegalArgumentException] { |
| 162 | + PageToken.fromRow(invalidRow) |
| 163 | + } |
| 164 | + assert(exception.getMessage.contains( |
| 165 | + "Invalid Page Token: required field")) |
| 166 | + } |
| 167 | +} |
0 commit comments