Search in sources :

Example 11 with CacheConfig

use of com.facebook.presto.cache.CacheConfig in project presto by prestodb.

the class TestHiveFileFormats method testDwrfOptimizedWriter.

@Test(dataProvider = "rowCount")
public void testDwrfOptimizedWriter(int rowCount) throws Exception {
    TestingConnectorSession session = new TestingConnectorSession(new HiveSessionProperties(new HiveClientConfig().setOrcOptimizedWriterEnabled(true).setOrcWriterValidationPercentage(100.0), new OrcFileWriterConfig(), new ParquetFileWriterConfig(), new CacheConfig()).getSessionProperties());
    // DWRF does not support modern Hive types
    // A Presto page can not contain a map with null keys, so a page based writer can not write null keys
    List<TestColumn> testColumns = TEST_COLUMNS.stream().filter(testColumn -> !hasType(testColumn.getObjectInspector(), PrimitiveCategory.DATE, PrimitiveCategory.VARCHAR, PrimitiveCategory.CHAR, PrimitiveCategory.DECIMAL)).filter(testColumn -> !testColumn.getName().equals("t_map_null_key") && !testColumn.getName().equals("t_map_null_key_complex_value") && !testColumn.getName().equals("t_map_null_key_complex_key_value")).collect(toList());
    assertThatFileFormat(DWRF).withColumns(testColumns).withRowsCount(rowCount).withSession(session).withFileWriterFactory(new OrcFileWriterFactory(HDFS_ENVIRONMENT, new OutputStreamDataSinkFactory(), FUNCTION_AND_TYPE_MANAGER, new NodeVersion("test"), HIVE_STORAGE_TIME_ZONE, STATS, new OrcFileWriterConfig(), NO_ENCRYPTION)).isReadableByRecordCursor(new GenericHiveRecordCursorProvider(HDFS_ENVIRONMENT)).isReadableByPageSource(new DwrfBatchPageSourceFactory(FUNCTION_AND_TYPE_MANAGER, FUNCTION_RESOLUTION, HIVE_CLIENT_CONFIG, HDFS_ENVIRONMENT, STATS, new StorageOrcFileTailSource(), StripeMetadataSourceFactory.of(new StorageStripeMetadataSource()), NO_ENCRYPTION));
}
Also used : RecordPageSource(com.facebook.presto.spi.RecordPageSource) DateTimeZone(org.joda.time.DateTimeZone) Arrays(java.util.Arrays) VarcharType.createUnboundedVarcharType(com.facebook.presto.common.type.VarcharType.createUnboundedVarcharType) PrimitiveObjectInspectorFactory.javaLongObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.PrimitiveObjectInspectorFactory.javaLongObjectInspector) Test(org.testng.annotations.Test) HIVE_PARTITION_SCHEMA_MISMATCH(com.facebook.presto.hive.HiveErrorCode.HIVE_PARTITION_SCHEMA_MISMATCH) PrimitiveObjectInspectorFactory.javaTimestampObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.PrimitiveObjectInspectorFactory.javaTimestampObjectInspector) FileSplit(org.apache.hadoop.mapred.FileSplit) Slices(io.airlift.slice.Slices) Configuration(org.apache.hadoop.conf.Configuration) Slices.utf8Slice(io.airlift.slice.Slices.utf8Slice) ObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.ObjectInspector) LzoCodec(io.airlift.compress.lzo.LzoCodec) DwrfBatchPageSourceFactory(com.facebook.presto.hive.orc.DwrfBatchPageSourceFactory) PrimitiveObjectInspectorFactory.javaFloatObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.PrimitiveObjectInspectorFactory.javaFloatObjectInspector) PrimitiveObjectInspectorFactory.javaDoubleObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.PrimitiveObjectInspectorFactory.javaDoubleObjectInspector) StorageFormat(com.facebook.presto.hive.metastore.StorageFormat) MapObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.MapObjectInspector) StructuralTestUtil.rowBlockOf(com.facebook.presto.tests.StructuralTestUtil.rowBlockOf) ImmutableList.toImmutableList(com.google.common.collect.ImmutableList.toImmutableList) ConnectorSession(com.facebook.presto.spi.ConnectorSession) INTEGER(com.facebook.presto.common.type.IntegerType.INTEGER) HIVE_CLIENT_CONFIG(com.facebook.presto.hive.HiveTestUtils.HIVE_CLIENT_CONFIG) Iterables.filter(com.google.common.collect.Iterables.filter) PrimitiveObjectInspectorFactory.javaIntObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.PrimitiveObjectInspectorFactory.javaIntObjectInspector) StructField(org.apache.hadoop.hive.serde2.objectinspector.StructField) MetadataReader(com.facebook.presto.parquet.cache.MetadataReader) ListObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.ListObjectInspector) PrimitiveCategory(org.apache.hadoop.hive.serde2.objectinspector.PrimitiveObjectInspector.PrimitiveCategory) FUNCTION_RESOLUTION(com.facebook.presto.hive.HiveTestUtils.FUNCTION_RESOLUTION) ROW_EXPRESSION_SERVICE(com.facebook.presto.hive.HiveTestUtils.ROW_EXPRESSION_SERVICE) Lists(com.google.common.collect.Lists) RCTEXT(com.facebook.presto.hive.HiveStorageFormat.RCTEXT) BOOLEAN(com.facebook.presto.common.type.BooleanType.BOOLEAN) CSV(com.facebook.presto.hive.HiveStorageFormat.CSV) ArrayType(com.facebook.presto.common.type.ArrayType) JSON(com.facebook.presto.hive.HiveStorageFormat.JSON) BIGINT(com.facebook.presto.common.type.BigintType.BIGINT) StorageStripeMetadataSource(com.facebook.presto.orc.StorageStripeMetadataSource) IOException(java.io.IOException) ObjectInspectorFactory.getStandardStructObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.ObjectInspectorFactory.getStandardStructObjectInspector) TestingConnectorSession(com.facebook.presto.testing.TestingConnectorSession) File(java.io.File) FUNCTION_AND_TYPE_MANAGER(com.facebook.presto.hive.HiveTestUtils.FUNCTION_AND_TYPE_MANAGER) RCBINARY(com.facebook.presto.hive.HiveStorageFormat.RCBINARY) HDFS_ENVIRONMENT(com.facebook.presto.hive.HiveTestUtils.HDFS_ENVIRONMENT) ParquetPageSourceFactory(com.facebook.presto.hive.parquet.ParquetPageSourceFactory) ParquetFileWriterFactory(com.facebook.presto.hive.parquet.ParquetFileWriterFactory) RowType(com.facebook.presto.common.type.RowType) ORC(com.facebook.presto.hive.HiveStorageFormat.ORC) CacheConfig(com.facebook.presto.cache.CacheConfig) SchemaTableName(com.facebook.presto.spi.SchemaTableName) AVRO(com.facebook.presto.hive.HiveStorageFormat.AVRO) TEXTFILE(com.facebook.presto.hive.HiveStorageFormat.TEXTFILE) StripeMetadataSourceFactory(com.facebook.presto.orc.StripeMetadataSourceFactory) ImmutableSet(com.google.common.collect.ImmutableSet) ImmutableMap(com.google.common.collect.ImmutableMap) DOUBLE(com.facebook.presto.common.type.DoubleType.DOUBLE) TimeZone(java.util.TimeZone) BeforeClass(org.testng.annotations.BeforeClass) DWRF(com.facebook.presto.hive.HiveStorageFormat.DWRF) Assert.assertNotNull(org.testng.Assert.assertNotNull) StructObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.StructObjectInspector) Instant(java.time.Instant) Collectors(java.util.stream.Collectors) TRUE_CONSTANT(com.facebook.presto.expressions.LogicalRowExpressions.TRUE_CONSTANT) RecordCursor(com.facebook.presto.spi.RecordCursor) List(java.util.List) StructuralTestUtil.arrayBlockOf(com.facebook.presto.tests.StructuralTestUtil.arrayBlockOf) VarcharTypeInfo(org.apache.hadoop.hive.serde2.typeinfo.VarcharTypeInfo) Optional(java.util.Optional) ErrorCodeSupplier(com.facebook.presto.spi.ErrorCodeSupplier) DataProvider(org.testng.annotations.DataProvider) Column(com.facebook.presto.hive.metastore.Column) PrimitiveObjectInspectorFactory.javaBooleanObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.PrimitiveObjectInspectorFactory.javaBooleanObjectInspector) HiveTestUtils.getTypes(com.facebook.presto.hive.HiveTestUtils.getTypes) HIVE_INVALID_PARTITION_VALUE(com.facebook.presto.hive.HiveErrorCode.HIVE_INVALID_PARTITION_VALUE) RcFilePageSourceFactory(com.facebook.presto.hive.rcfile.RcFilePageSourceFactory) Assert.assertEquals(org.testng.Assert.assertEquals) PrestoException(com.facebook.presto.spi.PrestoException) OptionalInt(java.util.OptionalInt) PrimitiveObjectInspectorFactory.getPrimitiveJavaObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.PrimitiveObjectInspectorFactory.getPrimitiveJavaObjectInspector) LzopCodec(io.airlift.compress.lzo.LzopCodec) PARQUET(com.facebook.presto.hive.HiveStorageFormat.PARQUET) NO_ENCRYPTION(com.facebook.presto.hive.HiveDwrfEncryptionProvider.NO_ENCRYPTION) StructuralTestUtil.mapBlockOf(com.facebook.presto.tests.StructuralTestUtil.mapBlockOf) HiveVarchar(org.apache.hadoop.hive.common.type.HiveVarchar) ImmutableList(com.google.common.collect.ImmutableList) PrimitiveObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.PrimitiveObjectInspector) SESSION(com.facebook.presto.hive.HiveTestUtils.SESSION) Objects.requireNonNull(java.util.Objects.requireNonNull) DEFAULT_HIVE_FILE_CONTEXT(com.facebook.presto.hive.HiveFileContext.DEFAULT_HIVE_FILE_CONTEXT) SEQUENCEFILE(com.facebook.presto.hive.HiveStorageFormat.SEQUENCEFILE) ObjectInspectorFactory.getStandardMapObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.ObjectInspectorFactory.getStandardMapObjectInspector) Storage(com.facebook.presto.hive.metastore.Storage) Assert.fail(org.testng.Assert.fail) OrcBatchPageSourceFactory(com.facebook.presto.hive.orc.OrcBatchPageSourceFactory) OutputStreamDataSinkFactory(com.facebook.presto.hive.datasink.OutputStreamDataSinkFactory) ObjectInspectorFactory.getStandardListObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.ObjectInspectorFactory.getStandardListObjectInspector) TupleDomain(com.facebook.presto.common.predicate.TupleDomain) StorageOrcFileTailSource(com.facebook.presto.orc.cache.StorageOrcFileTailSource) Collectors.toList(java.util.stream.Collectors.toList) ConnectorPageSource(com.facebook.presto.spi.ConnectorPageSource) Assert.assertTrue(org.testng.Assert.assertTrue) PrimitiveObjectInspectorFactory.javaStringObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.PrimitiveObjectInspectorFactory.javaStringObjectInspector) OutputStreamDataSinkFactory(com.facebook.presto.hive.datasink.OutputStreamDataSinkFactory) TestingConnectorSession(com.facebook.presto.testing.TestingConnectorSession) StorageOrcFileTailSource(com.facebook.presto.orc.cache.StorageOrcFileTailSource) DwrfBatchPageSourceFactory(com.facebook.presto.hive.orc.DwrfBatchPageSourceFactory) StorageStripeMetadataSource(com.facebook.presto.orc.StorageStripeMetadataSource) CacheConfig(com.facebook.presto.cache.CacheConfig) Test(org.testng.annotations.Test)

Example 12 with CacheConfig

use of com.facebook.presto.cache.CacheConfig in project presto by prestodb.

the class TestHiveFileFormats method testOrcOptimizedWriter.

@Test(dataProvider = "rowCount")
public void testOrcOptimizedWriter(int rowCount) throws Exception {
    TestingConnectorSession session = new TestingConnectorSession(new HiveSessionProperties(new HiveClientConfig().setOrcOptimizedWriterEnabled(true).setOrcWriterValidationPercentage(100.0), new OrcFileWriterConfig(), new ParquetFileWriterConfig(), new CacheConfig()).getSessionProperties());
    // A Presto page can not contain a map with null keys, so a page based writer can not write null keys
    List<TestColumn> testColumns = TEST_COLUMNS.stream().filter(testColumn -> !testColumn.getName().equals("t_map_null_key") && !testColumn.getName().equals("t_map_null_key_complex_value") && !testColumn.getName().equals("t_map_null_key_complex_key_value")).collect(toList());
    assertThatFileFormat(ORC).withColumns(testColumns).withRowsCount(rowCount).withSession(session).withFileWriterFactory(new OrcFileWriterFactory(HDFS_ENVIRONMENT, new OutputStreamDataSinkFactory(), FUNCTION_AND_TYPE_MANAGER, new NodeVersion("test"), HIVE_STORAGE_TIME_ZONE, STATS, new OrcFileWriterConfig(), NO_ENCRYPTION)).isReadableByRecordCursor(new GenericHiveRecordCursorProvider(HDFS_ENVIRONMENT)).isReadableByPageSource(new OrcBatchPageSourceFactory(FUNCTION_AND_TYPE_MANAGER, FUNCTION_RESOLUTION, false, HDFS_ENVIRONMENT, STATS, 100, new StorageOrcFileTailSource(), StripeMetadataSourceFactory.of(new StorageStripeMetadataSource())));
}
Also used : RecordPageSource(com.facebook.presto.spi.RecordPageSource) DateTimeZone(org.joda.time.DateTimeZone) Arrays(java.util.Arrays) VarcharType.createUnboundedVarcharType(com.facebook.presto.common.type.VarcharType.createUnboundedVarcharType) PrimitiveObjectInspectorFactory.javaLongObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.PrimitiveObjectInspectorFactory.javaLongObjectInspector) Test(org.testng.annotations.Test) HIVE_PARTITION_SCHEMA_MISMATCH(com.facebook.presto.hive.HiveErrorCode.HIVE_PARTITION_SCHEMA_MISMATCH) PrimitiveObjectInspectorFactory.javaTimestampObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.PrimitiveObjectInspectorFactory.javaTimestampObjectInspector) FileSplit(org.apache.hadoop.mapred.FileSplit) Slices(io.airlift.slice.Slices) Configuration(org.apache.hadoop.conf.Configuration) Slices.utf8Slice(io.airlift.slice.Slices.utf8Slice) ObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.ObjectInspector) LzoCodec(io.airlift.compress.lzo.LzoCodec) DwrfBatchPageSourceFactory(com.facebook.presto.hive.orc.DwrfBatchPageSourceFactory) PrimitiveObjectInspectorFactory.javaFloatObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.PrimitiveObjectInspectorFactory.javaFloatObjectInspector) PrimitiveObjectInspectorFactory.javaDoubleObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.PrimitiveObjectInspectorFactory.javaDoubleObjectInspector) StorageFormat(com.facebook.presto.hive.metastore.StorageFormat) MapObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.MapObjectInspector) StructuralTestUtil.rowBlockOf(com.facebook.presto.tests.StructuralTestUtil.rowBlockOf) ImmutableList.toImmutableList(com.google.common.collect.ImmutableList.toImmutableList) ConnectorSession(com.facebook.presto.spi.ConnectorSession) INTEGER(com.facebook.presto.common.type.IntegerType.INTEGER) HIVE_CLIENT_CONFIG(com.facebook.presto.hive.HiveTestUtils.HIVE_CLIENT_CONFIG) Iterables.filter(com.google.common.collect.Iterables.filter) PrimitiveObjectInspectorFactory.javaIntObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.PrimitiveObjectInspectorFactory.javaIntObjectInspector) StructField(org.apache.hadoop.hive.serde2.objectinspector.StructField) MetadataReader(com.facebook.presto.parquet.cache.MetadataReader) ListObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.ListObjectInspector) PrimitiveCategory(org.apache.hadoop.hive.serde2.objectinspector.PrimitiveObjectInspector.PrimitiveCategory) FUNCTION_RESOLUTION(com.facebook.presto.hive.HiveTestUtils.FUNCTION_RESOLUTION) ROW_EXPRESSION_SERVICE(com.facebook.presto.hive.HiveTestUtils.ROW_EXPRESSION_SERVICE) Lists(com.google.common.collect.Lists) RCTEXT(com.facebook.presto.hive.HiveStorageFormat.RCTEXT) BOOLEAN(com.facebook.presto.common.type.BooleanType.BOOLEAN) CSV(com.facebook.presto.hive.HiveStorageFormat.CSV) ArrayType(com.facebook.presto.common.type.ArrayType) JSON(com.facebook.presto.hive.HiveStorageFormat.JSON) BIGINT(com.facebook.presto.common.type.BigintType.BIGINT) StorageStripeMetadataSource(com.facebook.presto.orc.StorageStripeMetadataSource) IOException(java.io.IOException) ObjectInspectorFactory.getStandardStructObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.ObjectInspectorFactory.getStandardStructObjectInspector) TestingConnectorSession(com.facebook.presto.testing.TestingConnectorSession) File(java.io.File) FUNCTION_AND_TYPE_MANAGER(com.facebook.presto.hive.HiveTestUtils.FUNCTION_AND_TYPE_MANAGER) RCBINARY(com.facebook.presto.hive.HiveStorageFormat.RCBINARY) HDFS_ENVIRONMENT(com.facebook.presto.hive.HiveTestUtils.HDFS_ENVIRONMENT) ParquetPageSourceFactory(com.facebook.presto.hive.parquet.ParquetPageSourceFactory) ParquetFileWriterFactory(com.facebook.presto.hive.parquet.ParquetFileWriterFactory) RowType(com.facebook.presto.common.type.RowType) ORC(com.facebook.presto.hive.HiveStorageFormat.ORC) CacheConfig(com.facebook.presto.cache.CacheConfig) SchemaTableName(com.facebook.presto.spi.SchemaTableName) AVRO(com.facebook.presto.hive.HiveStorageFormat.AVRO) TEXTFILE(com.facebook.presto.hive.HiveStorageFormat.TEXTFILE) StripeMetadataSourceFactory(com.facebook.presto.orc.StripeMetadataSourceFactory) ImmutableSet(com.google.common.collect.ImmutableSet) ImmutableMap(com.google.common.collect.ImmutableMap) DOUBLE(com.facebook.presto.common.type.DoubleType.DOUBLE) TimeZone(java.util.TimeZone) BeforeClass(org.testng.annotations.BeforeClass) DWRF(com.facebook.presto.hive.HiveStorageFormat.DWRF) Assert.assertNotNull(org.testng.Assert.assertNotNull) StructObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.StructObjectInspector) Instant(java.time.Instant) Collectors(java.util.stream.Collectors) TRUE_CONSTANT(com.facebook.presto.expressions.LogicalRowExpressions.TRUE_CONSTANT) RecordCursor(com.facebook.presto.spi.RecordCursor) List(java.util.List) StructuralTestUtil.arrayBlockOf(com.facebook.presto.tests.StructuralTestUtil.arrayBlockOf) VarcharTypeInfo(org.apache.hadoop.hive.serde2.typeinfo.VarcharTypeInfo) Optional(java.util.Optional) ErrorCodeSupplier(com.facebook.presto.spi.ErrorCodeSupplier) DataProvider(org.testng.annotations.DataProvider) Column(com.facebook.presto.hive.metastore.Column) PrimitiveObjectInspectorFactory.javaBooleanObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.PrimitiveObjectInspectorFactory.javaBooleanObjectInspector) HiveTestUtils.getTypes(com.facebook.presto.hive.HiveTestUtils.getTypes) HIVE_INVALID_PARTITION_VALUE(com.facebook.presto.hive.HiveErrorCode.HIVE_INVALID_PARTITION_VALUE) RcFilePageSourceFactory(com.facebook.presto.hive.rcfile.RcFilePageSourceFactory) Assert.assertEquals(org.testng.Assert.assertEquals) PrestoException(com.facebook.presto.spi.PrestoException) OptionalInt(java.util.OptionalInt) PrimitiveObjectInspectorFactory.getPrimitiveJavaObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.PrimitiveObjectInspectorFactory.getPrimitiveJavaObjectInspector) LzopCodec(io.airlift.compress.lzo.LzopCodec) PARQUET(com.facebook.presto.hive.HiveStorageFormat.PARQUET) NO_ENCRYPTION(com.facebook.presto.hive.HiveDwrfEncryptionProvider.NO_ENCRYPTION) StructuralTestUtil.mapBlockOf(com.facebook.presto.tests.StructuralTestUtil.mapBlockOf) HiveVarchar(org.apache.hadoop.hive.common.type.HiveVarchar) ImmutableList(com.google.common.collect.ImmutableList) PrimitiveObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.PrimitiveObjectInspector) SESSION(com.facebook.presto.hive.HiveTestUtils.SESSION) Objects.requireNonNull(java.util.Objects.requireNonNull) DEFAULT_HIVE_FILE_CONTEXT(com.facebook.presto.hive.HiveFileContext.DEFAULT_HIVE_FILE_CONTEXT) SEQUENCEFILE(com.facebook.presto.hive.HiveStorageFormat.SEQUENCEFILE) ObjectInspectorFactory.getStandardMapObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.ObjectInspectorFactory.getStandardMapObjectInspector) Storage(com.facebook.presto.hive.metastore.Storage) Assert.fail(org.testng.Assert.fail) OrcBatchPageSourceFactory(com.facebook.presto.hive.orc.OrcBatchPageSourceFactory) OutputStreamDataSinkFactory(com.facebook.presto.hive.datasink.OutputStreamDataSinkFactory) ObjectInspectorFactory.getStandardListObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.ObjectInspectorFactory.getStandardListObjectInspector) TupleDomain(com.facebook.presto.common.predicate.TupleDomain) StorageOrcFileTailSource(com.facebook.presto.orc.cache.StorageOrcFileTailSource) Collectors.toList(java.util.stream.Collectors.toList) ConnectorPageSource(com.facebook.presto.spi.ConnectorPageSource) Assert.assertTrue(org.testng.Assert.assertTrue) PrimitiveObjectInspectorFactory.javaStringObjectInspector(org.apache.hadoop.hive.serde2.objectinspector.primitive.PrimitiveObjectInspectorFactory.javaStringObjectInspector) OutputStreamDataSinkFactory(com.facebook.presto.hive.datasink.OutputStreamDataSinkFactory) OrcBatchPageSourceFactory(com.facebook.presto.hive.orc.OrcBatchPageSourceFactory) TestingConnectorSession(com.facebook.presto.testing.TestingConnectorSession) StorageOrcFileTailSource(com.facebook.presto.orc.cache.StorageOrcFileTailSource) StorageStripeMetadataSource(com.facebook.presto.orc.StorageStripeMetadataSource) CacheConfig(com.facebook.presto.cache.CacheConfig) Test(org.testng.annotations.Test)

Example 13 with CacheConfig

use of com.facebook.presto.cache.CacheConfig in project presto by prestodb.

the class TestHiveFileFormats method testOrcUseColumnNames.

@Test(dataProvider = "rowCount")
public void testOrcUseColumnNames(int rowCount) throws Exception {
    TestingConnectorSession session = new TestingConnectorSession(new HiveSessionProperties(new HiveClientConfig(), new OrcFileWriterConfig(), new ParquetFileWriterConfig(), new CacheConfig()).getSessionProperties());
    assertThatFileFormat(ORC).withWriteColumns(TEST_COLUMNS).withRowsCount(rowCount).withReadColumns(Lists.reverse(TEST_COLUMNS)).withSession(session).isReadableByPageSource(new OrcBatchPageSourceFactory(FUNCTION_AND_TYPE_MANAGER, FUNCTION_RESOLUTION, true, HDFS_ENVIRONMENT, STATS, 100, new StorageOrcFileTailSource(), StripeMetadataSourceFactory.of(new StorageStripeMetadataSource())));
}
Also used : OrcBatchPageSourceFactory(com.facebook.presto.hive.orc.OrcBatchPageSourceFactory) TestingConnectorSession(com.facebook.presto.testing.TestingConnectorSession) StorageOrcFileTailSource(com.facebook.presto.orc.cache.StorageOrcFileTailSource) StorageStripeMetadataSource(com.facebook.presto.orc.StorageStripeMetadataSource) CacheConfig(com.facebook.presto.cache.CacheConfig) Test(org.testng.annotations.Test)

Example 14 with CacheConfig

use of com.facebook.presto.cache.CacheConfig in project presto by prestodb.

the class TestAlluxioCachingFileSystem method testBasic.

private void testBasic(boolean validationEnabled) throws Exception {
    CacheConfig cacheConfig = new CacheConfig().setCacheType(ALLUXIO).setCachingEnabled(true).setBaseDirectory(cacheDirectory).setValidationEnabled(validationEnabled);
    AlluxioCacheConfig alluxioCacheConfig = new AlluxioCacheConfig();
    Configuration configuration = getHdfsConfiguration(cacheConfig, alluxioCacheConfig);
    AlluxioCachingFileSystem fileSystem = cachingFileSystem(configuration);
    Path p = new Path("/tmp");
    assertEquals(fileSystem.getDefaultBlockSize(p), 1024L);
    assertEquals(fileSystem.getDefaultReplication(p), 10);
    byte[] buffer = new byte[PAGE_SIZE * 2];
    int pageOffset = PAGE_SIZE;
    // new read
    resetBaseline();
    assertEquals(readFully(fileSystem, pageOffset + 10, buffer, 0, 100), 100);
    checkMetrics(MetricKey.CLIENT_CACHE_BYTES_READ_CACHE, 0);
    checkMetrics(MetricKey.CLIENT_CACHE_BYTES_REQUESTED_EXTERNAL, 100);
    checkMetrics(MetricKey.CLIENT_CACHE_BYTES_READ_EXTERNAL, PAGE_SIZE);
    checkMetrics(MetricKey.CLIENT_CACHE_BYTES_EVICTED, 0);
    validateBuffer(data, pageOffset + 10, buffer, 0, 100);
    // read within the cached page
    resetBaseline();
    assertEquals(readFully(fileSystem, pageOffset + 20, buffer, 0, 90), 90);
    checkMetrics(MetricKey.CLIENT_CACHE_BYTES_READ_CACHE, 90);
    checkMetrics(MetricKey.CLIENT_CACHE_BYTES_REQUESTED_EXTERNAL, 0);
    checkMetrics(MetricKey.CLIENT_CACHE_BYTES_READ_EXTERNAL, 0);
    checkMetrics(MetricKey.CLIENT_CACHE_BYTES_EVICTED, 0);
    validateBuffer(data, pageOffset + 20, buffer, 0, 90);
    // read partially after the range of the cache
    resetBaseline();
    assertEquals(readFully(fileSystem, pageOffset + PAGE_SIZE - 10, buffer, 0, 100), 100);
    checkMetrics(MetricKey.CLIENT_CACHE_BYTES_READ_CACHE, 10);
    checkMetrics(MetricKey.CLIENT_CACHE_BYTES_REQUESTED_EXTERNAL, 90);
    checkMetrics(MetricKey.CLIENT_CACHE_BYTES_READ_EXTERNAL, PAGE_SIZE);
    checkMetrics(MetricKey.CLIENT_CACHE_BYTES_EVICTED, 0);
    validateBuffer(data, pageOffset + PAGE_SIZE - 10, buffer, 0, 100);
    // read partially before the range of the cache
    resetBaseline();
    assertEquals(readFully(fileSystem, pageOffset - 10, buffer, 10, 50), 50);
    checkMetrics(MetricKey.CLIENT_CACHE_BYTES_READ_CACHE, 40);
    checkMetrics(MetricKey.CLIENT_CACHE_BYTES_REQUESTED_EXTERNAL, 10);
    checkMetrics(MetricKey.CLIENT_CACHE_BYTES_READ_EXTERNAL, PAGE_SIZE);
    checkMetrics(MetricKey.CLIENT_CACHE_BYTES_EVICTED, 0);
    validateBuffer(data, pageOffset - 10, buffer, 10, 50);
    // skip one page
    resetBaseline();
    assertEquals(readFully(fileSystem, pageOffset + PAGE_SIZE * 3, buffer, 40, 50), 50);
    checkMetrics(MetricKey.CLIENT_CACHE_BYTES_READ_CACHE, 0);
    checkMetrics(MetricKey.CLIENT_CACHE_BYTES_REQUESTED_EXTERNAL, 50);
    checkMetrics(MetricKey.CLIENT_CACHE_BYTES_READ_EXTERNAL, PAGE_SIZE);
    checkMetrics(MetricKey.CLIENT_CACHE_BYTES_EVICTED, 0);
    validateBuffer(data, pageOffset + PAGE_SIZE * 3, buffer, 40, 50);
    // read between cached pages
    resetBaseline();
    assertEquals(readFully(fileSystem, pageOffset + PAGE_SIZE * 2 - 10, buffer, 400, PAGE_SIZE + 20), PAGE_SIZE + 20);
    checkMetrics(MetricKey.CLIENT_CACHE_BYTES_READ_CACHE, 20);
    checkMetrics(MetricKey.CLIENT_CACHE_BYTES_REQUESTED_EXTERNAL, PAGE_SIZE);
    checkMetrics(MetricKey.CLIENT_CACHE_BYTES_READ_EXTERNAL, PAGE_SIZE);
    checkMetrics(MetricKey.CLIENT_CACHE_BYTES_EVICTED, 0);
    validateBuffer(data, pageOffset + PAGE_SIZE * 2 - 10, buffer, 400, PAGE_SIZE + 20);
}
Also used : Path(org.apache.hadoop.fs.Path) Configuration(org.apache.hadoop.conf.Configuration) CacheConfig(com.facebook.presto.cache.CacheConfig)

Example 15 with CacheConfig

use of com.facebook.presto.cache.CacheConfig in project presto by prestodb.

the class TestAlluxioCachingFileSystem method testSyncRestoreFailure.

@Test(timeOut = 30_000, expectedExceptions = { IOException.class })
public void testSyncRestoreFailure() throws Exception {
    URI badCacheDirectory = createTempDirectory("alluxio_cache_bad").toUri();
    File cacheDirectory = new File(badCacheDirectory.getPath());
    cacheDirectory.setWritable(false);
    CacheConfig cacheConfig = new CacheConfig().setCacheType(ALLUXIO).setCachingEnabled(true).setBaseDirectory(badCacheDirectory);
    AlluxioCacheConfig alluxioCacheConfig = new AlluxioCacheConfig();
    Configuration configuration = getHdfsConfiguration(cacheConfig, alluxioCacheConfig);
    try {
        cachingFileSystem(configuration);
    } finally {
        cacheDirectory.setWritable(true);
    }
}
Also used : Configuration(org.apache.hadoop.conf.Configuration) URI(java.net.URI) File(java.io.File) CacheConfig(com.facebook.presto.cache.CacheConfig) Test(org.testng.annotations.Test) TestingCacheUtils.stressTest(com.facebook.presto.cache.TestingCacheUtils.stressTest)

Aggregations

CacheConfig (com.facebook.presto.cache.CacheConfig)45 Test (org.testng.annotations.Test)33 TestingConnectorSession (com.facebook.presto.testing.TestingConnectorSession)31 ConnectorSession (com.facebook.presto.spi.ConnectorSession)18 Configuration (org.apache.hadoop.conf.Configuration)12 SchemaTableName (com.facebook.presto.spi.SchemaTableName)9 ConnectorPageSource (com.facebook.presto.spi.ConnectorPageSource)8 HiveClientConfig (com.facebook.presto.hive.HiveClientConfig)7 HiveSessionProperties (com.facebook.presto.hive.HiveSessionProperties)7 OrcFileWriterConfig (com.facebook.presto.hive.OrcFileWriterConfig)7 ParquetFileWriterConfig (com.facebook.presto.hive.ParquetFileWriterConfig)7 TestingCacheUtils.stressTest (com.facebook.presto.cache.TestingCacheUtils.stressTest)6 Storage (com.facebook.presto.hive.metastore.Storage)6 File (java.io.File)6 StorageFormat (com.facebook.presto.hive.metastore.StorageFormat)5 OutputStreamDataSinkFactory (com.facebook.presto.hive.datasink.OutputStreamDataSinkFactory)4 RcFilePageSourceFactory (com.facebook.presto.hive.rcfile.RcFilePageSourceFactory)4 DataSize (io.airlift.units.DataSize)4 BeforeClass (org.testng.annotations.BeforeClass)4 CounterStat (com.facebook.airlift.stats.CounterStat)3