pvary commented on code in PR #17610: URL: https://github.com/apache/iceberg/pull/17610#discussion_r4454138031
########## spark/v4.2/spark/src/test/java/org/apache/iceberg/spark/data/TestSparkVectorizedFormatModel.java: ########## @@ -0,0 +1,209 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one + * or more contributor license agreements. See the NOTICE file + * distributed with this work for additional information + * regarding copyright ownership. The ASF licenses this file + * to you under the Apache License, Version 2.0 (the + * "License"); you may not use this file except in compliance + * with the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, + * software distributed under the License is distributed on an + * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY + * KIND, either express or implied. See the License for the + * specific language governing permissions and limitations + * under the License. + */ +package org.apache.iceberg.spark.data; + +import static org.assertj.core.api.Assertions.assertThat; +import static org.assertj.core.api.Assumptions.assumeThat; + +import java.io.IOException; +import java.util.Arrays; +import java.util.Collection; +import java.util.List; +import java.util.Set; +import org.apache.arrow.vector.ValueVector; +import org.apache.iceberg.FileFormat; +import org.apache.iceberg.Schema; +import org.apache.iceberg.data.DataGenerator; +import org.apache.iceberg.data.DataGenerators; +import org.apache.iceberg.data.ReadFormatModelTests; +import org.apache.iceberg.data.Record; +import org.apache.iceberg.io.CloseableIterable; +import org.apache.iceberg.spark.SparkSchemaUtil; +import org.apache.iceberg.spark.SparkUtil; +import org.apache.iceberg.spark.data.vectorized.IcebergArrowColumnVector; +import org.apache.iceberg.types.Type; +import org.apache.spark.sql.vectorized.ColumnarBatch; +import org.junit.jupiter.params.ParameterizedTest; +import org.junit.jupiter.params.provider.Arguments; +import org.junit.jupiter.params.provider.FieldSource; + +class TestSparkVectorizedFormatModel extends ReadFormatModelTests<ColumnarBatch> { + + private static final FileFormat[] FILE_FORMATS = {FileFormat.ORC, FileFormat.PARQUET}; + + private static final Set<Type.TypeID> SPARK_UNSUPPORTED_TYPE_IDS = + Set.of(Type.TypeID.TIME, Type.TypeID.TIMESTAMP_NANO); + + // The Arrow-based Parquet reader cannot read nested or geospatial columns; + // the vectorized ORC reader supports nested types. + private static final Set<Type.TypeID> ARROW_UNSUPPORTED_TYPE_IDS = + Set.of( + Type.TypeID.STRUCT, + Type.TypeID.LIST, + Type.TypeID.MAP, + Type.TypeID.GEOMETRY, + Type.TypeID.GEOGRAPHY); + + private static final Set<Type.TypeID> PARQUET_UNSUPPORTED_TYPE_IDS = + Set.of( + Type.TypeID.TIME, + Type.TypeID.TIMESTAMP_NANO, + Type.TypeID.STRUCT, + Type.TypeID.LIST, + Type.TypeID.MAP, + Type.TypeID.GEOMETRY, + Type.TypeID.GEOGRAPHY); + + private static final List<Arguments> FORMAT_AND_GENERATOR = + Arrays.stream(FILE_FORMATS) + .flatMap( + format -> + Arrays.stream(DataGenerators.ALL) + .filter(generator -> supportsVectorized(format, generator)) + .map(generator -> Arguments.of(format, generator))) + .toList(); + + private static boolean supportsVectorized(FileFormat format, DataGenerator generator) { + return supportsGenerator(format, generator) + && (format != FileFormat.PARQUET + || !containsType(generator.schema(), ARROW_UNSUPPORTED_TYPE_IDS)); + } + + @Override + protected Collection<Type.TypeID> unsupportedTypeIds(FileFormat fileFormat) { + return fileFormat == FileFormat.PARQUET + ? PARQUET_UNSUPPORTED_TYPE_IDS + : SPARK_UNSUPPORTED_TYPE_IDS; + } + + @Override + protected boolean supportsBatchReads() { + return true; + } + + @Override + protected int countRows(List<ColumnarBatch> batches) { + return batches.stream().mapToInt(ColumnarBatch::numRows).sum(); + } + + @Override + protected void assertContainersReused(ColumnarBatch first, ColumnarBatch second) { + assertThat(second.numCols()).isEqualTo(first.numCols()); + + for (int i = 0; i < first.numCols(); i++) { + assertThat(arrowVector(second, i)).isSameAs(arrowVector(first, i)); + } + } + + @Override + @ParameterizedTest + @FieldSource("FILE_FORMATS") + protected void testReadMetadataColumnIsDeleted(FileFormat fileFormat) throws IOException { + assumeThat(fileFormat) + .as("Vectorized Parquet reader only populates _deleted when a delete filter is set") + .isNotEqualTo(FileFormat.PARQUET); + super.testReadMetadataColumnIsDeleted(fileFormat); + } + + @Override + @ParameterizedTest + @FieldSource("FILE_FORMATS") + protected void testReadMetadataColumnRowLineage(FileFormat fileFormat) throws IOException { + assumeRowLineageSupported(fileFormat); + super.testReadMetadataColumnRowLineage(fileFormat); + } + + @Override + @ParameterizedTest + @FieldSource("FILE_FORMATS") + protected void testReadMetadataColumnRowLineageExistingValues(FileFormat fileFormat) + throws IOException { + assumeRowLineageSupported(fileFormat); + super.testReadMetadataColumnRowLineageExistingValues(fileFormat); + } + + // TODO: remove when https://github.com/apache/iceberg/issues/18543 is fixed + private static void assumeRowLineageSupported(FileFormat fileFormat) { + assumeThat(fileFormat) + .as("Vectorized ORC reader does not apply row lineage") + .isNotEqualTo(FileFormat.ORC); + } + + private ValueVector arrowVector(ColumnarBatch batch, int col) { + assertThat(batch.column(col)).isInstanceOf(IcebergArrowColumnVector.class); + + IcebergArrowColumnVector vector = (IcebergArrowColumnVector) batch.column(col); + return vector.vectorAccessor().getVector(); + } + + @Override + protected void assertContainersNotReused(ColumnarBatch first, ColumnarBatch second) { + assertThat(second.numCols()).isEqualTo(first.numCols()); + + for (int i = 0; i < first.numCols(); i++) { + assertThat(arrowVector(second, i)).isNotSameAs(arrowVector(first, i)); + } + } + + @Override + protected Object engineSchema(Schema schema) { + return SparkSchemaUtil.convert(schema); + } + + @Override + protected Object convertConstantToEngine(Type type, Object value) { + return SparkUtil.internalToSpark(type, value); + } + + @Override + protected Class<ColumnarBatch> engineType() { + return ColumnarBatch.class; + } + + @Override + protected ColumnarBatch convertToEngine(Record record, Schema schema) { + throw new UnsupportedOperationException("Vectorized reads are read-only"); + } + + @Override + protected void assertEquals( + Schema schema, List<ColumnarBatch> expected, List<ColumnarBatch> actual) { + throw new UnsupportedOperationException("Use assertRecordsEqual for batch reads"); + } + + @Override + protected void assertRecordsEqual( + Schema schema, List<Record> expected, CloseableIterable<ColumnarBatch> actual) { + int rowOffset = 0; + + for (ColumnarBatch batch : actual) { + assertThat(rowOffset + batch.numRows()).isLessThanOrEqualTo(expected.size()); + + for (int i = 0; i < batch.numRows(); i++) { + TestHelpers.assertEquals( + schema, + InternalRowConverter.convert(schema, expected.get(rowOffset + i)), + batch.getRow(i)); + } + rowOffset += batch.numRows(); Review Comment: nit: newline -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
