pvary commented on code in PR #15906:
URL: https://github.com/apache/iceberg/pull/15906#discussion_r3086680647
##########
data/src/test/java/org/apache/iceberg/data/BaseFormatModelTests.java:
##########
@@ -628,6 +639,326 @@ void
testReaderBuilderRecordsPerBatchNotSupported(FileFormat fileFormat) throws
.isInstanceOf(UnsupportedOperationException.class);
}
+ @ParameterizedTest
+ @FieldSource("FILE_FORMATS")
+ void testDataWriterMetricsCollection(FileFormat fileFormat) throws
IOException {
+ DataGenerator dataGenerator = new DataGenerators.DefaultSchema();
+ Schema schema = dataGenerator.schema();
+ FileWriterBuilder<DataWriter<Record>, Object> writerBuilder =
+ FormatModelRegistry.dataWriteBuilder(fileFormat, Record.class,
encryptedFile);
+
+ DataWriter<Record> writer =
+
writerBuilder.schema(schema).spec(PartitionSpec.unpartitioned()).build();
+
+ List<Record> genericRecords = dataGenerator.generateRecords();
+
+ try (writer) {
+ genericRecords.forEach(writer::write);
+ }
+
+ DataFile dataFile = writer.toDataFile();
+
+ assertThat(dataFile).isNotNull();
+ assertThat(dataFile.recordCount()).isEqualTo(genericRecords.size());
+
+ // AVRO Only support metric collection with rowCount
+ assumeSupports(fileFormat, FEATURE_METRIC_COLLECT);
+
+ assertValueAndNullCounts(
+ schema, genericRecords, dataFile.valueCounts(),
dataFile.nullValueCounts());
+ assertBounds(schema, genericRecords, dataFile.lowerBounds(),
dataFile.upperBounds());
+
+ assertThat(dataFile.columnSizes()).isNotNull().isNotEmpty();
+ }
+
+ @ParameterizedTest
+ @FieldSource("FILE_FORMATS")
+ void testDataWriterMetricsWithNoneMode(FileFormat fileFormat) throws
IOException {
+
+ // AVRO Only support metric collection with rowCount
+ assumeSupports(fileFormat, FEATURE_METRIC_COLLECT);
+
+ DataGenerator dataGenerator = new DataGenerators.DefaultSchema();
+ Schema schema = dataGenerator.schema();
+
+ TestTables.TestTable table =
+ TestTables.create(
+ tableDir,
+ "test",
+ schema,
+ PartitionSpec.unpartitioned(),
+ 3,
+ ImmutableMap.of(TableProperties.DEFAULT_WRITE_METRICS_MODE,
"none"));
+
+ MetricsConfig noneConfig = MetricsConfig.forTable(table);
+
+ FileWriterBuilder<DataWriter<Record>, Object> writerBuilder =
+ FormatModelRegistry.dataWriteBuilder(fileFormat, Record.class,
encryptedFile);
+
+ DataWriter<Record> writer =
+ writerBuilder
+ .schema(schema)
+ .spec(PartitionSpec.unpartitioned())
+ .metricsConfig(noneConfig)
+ .build();
+
+ List<Record> genericRecords = dataGenerator.generateRecords();
+
+ try (writer) {
+ genericRecords.forEach(writer::write);
+ }
+
+ DataFile dataFile = writer.toDataFile();
+
+ assertThat(dataFile).isNotNull();
+ assertThat(dataFile.recordCount()).isEqualTo(genericRecords.size());
+
+ // In `None` mode, `valueCounts`, `nullValueCounts`, `lowerBounds`, and
`upperBounds` should be
+ // empty/null.
+ for (Types.NestedField field : schema.columns()) {
+ if (field.type().isPrimitiveType()) {
+ assertThat(dataFile.valueCounts().get(field.fieldId())).isNull();
+ assertThat(dataFile.nullValueCounts().get(field.fieldId())).isNull();
+ assertThat(dataFile.lowerBounds().get(field.fieldId())).isNull();
+ assertThat(dataFile.upperBounds().get(field.fieldId())).isNull();
+ }
+ }
+
+ TestTables.clearTables();
+ }
+
+ @ParameterizedTest
+ @FieldSource("FILE_FORMATS")
+ void testDataWriterMetricsWithCountsMode(FileFormat fileFormat) throws
IOException {
+ // AVRO Only support metric collection with rowCount
+ assumeSupports(fileFormat, FEATURE_METRIC_COLLECT);
+
+ DataGenerator dataGenerator = new DataGenerators.DefaultSchema();
+ Schema schema = dataGenerator.schema();
+ TestTables.TestTable table =
+ TestTables.create(
+ tableDir,
+ "test",
+ schema,
+ PartitionSpec.unpartitioned(),
+ 3,
+ ImmutableMap.of(TableProperties.DEFAULT_WRITE_METRICS_MODE,
"counts"));
+
+ MetricsConfig countsConfig = MetricsConfig.forTable(table);
+
+ FileWriterBuilder<DataWriter<Record>, Object> writerBuilder =
+ FormatModelRegistry.dataWriteBuilder(fileFormat, Record.class,
encryptedFile);
+
+ DataWriter<Record> writer =
+ writerBuilder
+ .schema(schema)
+ .spec(PartitionSpec.unpartitioned())
+ .metricsConfig(countsConfig)
+ .build();
+
+ List<Record> genericRecords = dataGenerator.generateRecords();
+
+ try (writer) {
+ genericRecords.forEach(writer::write);
+ }
+
+ DataFile dataFile = writer.toDataFile();
+
+ assertThat(dataFile).isNotNull();
+ assertThat(dataFile.recordCount()).isEqualTo(genericRecords.size());
+
+ // In the counts mode, valueCounts and nullValueCounts should be present,
while lowerBounds and
+ // upperBounds should be null.
+ assertValueAndNullCounts(
+ schema, genericRecords, dataFile.valueCounts(),
dataFile.nullValueCounts());
+ assertBoundsNull(schema, dataFile.lowerBounds(), dataFile.upperBounds());
+
+ TestTables.clearTables();
+ }
+
+ @ParameterizedTest
+ @FieldSource("FILE_FORMATS")
+ void testEqualityDeleteWriterMetricsCollection(FileFormat fileFormat) throws
IOException {
+
+ // AVRO Only support metric collection with rowCount
+ assumeSupports(fileFormat, FEATURE_METRIC_COLLECT);
+
+ DataGenerator dataGenerator = new DataGenerators.DefaultSchema();
+ Schema schema = dataGenerator.schema();
+ FileWriterBuilder<EqualityDeleteWriter<Record>, Object> writerBuilder =
+ FormatModelRegistry.equalityDeleteWriteBuilder(fileFormat,
Record.class, encryptedFile);
+
+ EqualityDeleteWriter<Record> writer =
+ writerBuilder
+ .schema(schema)
+ .spec(PartitionSpec.unpartitioned())
+ .equalityFieldIds(1)
+ .build();
+
+ List<Record> genericRecords = dataGenerator.generateRecords();
+
+ try (writer) {
+ genericRecords.forEach(writer::write);
+ }
+
+ DeleteFile deleteFile = writer.toDeleteFile();
+
+ assertThat(deleteFile).isNotNull();
+ assertThat(deleteFile.recordCount()).isEqualTo(genericRecords.size());
+
+ assertValueAndNullCounts(
+ schema, genericRecords, deleteFile.valueCounts(),
deleteFile.nullValueCounts());
+ assertBounds(schema, genericRecords, deleteFile.lowerBounds(),
deleteFile.upperBounds());
+ }
+
+ @ParameterizedTest
+ @FieldSource("FILE_FORMATS")
+ void testPositionDeleteWriterMetricsSingleFile(FileFormat fileFormat) throws
IOException {
+ Schema positionDeleteSchema = DeleteSchemaUtil.pathPosSchema();
Review Comment:
How much of these methods are duplicated? Can we refactor them to have a
cleaner code?
--
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]
