Search in sources :

Example 1 with ProjectionApplicationResult

use of io.trino.spi.connector.ProjectionApplicationResult in project trino by trinodb.

the class MetadataManager method applyProjection.

@Override
public Optional<ProjectionApplicationResult<TableHandle>> applyProjection(Session session, TableHandle table, List<ConnectorExpression> projections, Map<String, ColumnHandle> assignments) {
    CatalogName catalogName = table.getCatalogName();
    ConnectorMetadata metadata = getMetadata(session, catalogName);
    ConnectorSession connectorSession = session.toConnectorSession(catalogName);
    return metadata.applyProjection(connectorSession, table.getConnectorHandle(), projections, assignments).map(result -> {
        verifyProjection(table, result.getProjections(), result.getAssignments(), projections.size());
        return new ProjectionApplicationResult<>(new TableHandle(catalogName, result.getHandle(), table.getTransaction()), result.getProjections(), result.getAssignments(), result.isPrecalculateStatistics());
    });
}
Also used : CatalogName(io.trino.connector.CatalogName) ConnectorSession(io.trino.spi.connector.ConnectorSession) ConnectorOutputTableHandle(io.trino.spi.connector.ConnectorOutputTableHandle) ConnectorTableHandle(io.trino.spi.connector.ConnectorTableHandle) ConnectorInsertTableHandle(io.trino.spi.connector.ConnectorInsertTableHandle) ConnectorMetadata(io.trino.spi.connector.ConnectorMetadata) ProjectionApplicationResult(io.trino.spi.connector.ProjectionApplicationResult)

Example 2 with ProjectionApplicationResult

use of io.trino.spi.connector.ProjectionApplicationResult in project trino by trinodb.

the class AbstractTestHive method testApplyProjection.

@Test
public void testApplyProjection() throws Exception {
    ColumnMetadata bigIntColumn0 = new ColumnMetadata("int0", BIGINT);
    ColumnMetadata bigIntColumn1 = new ColumnMetadata("int1", BIGINT);
    RowType oneLevelRowType = toRowType(ImmutableList.of(bigIntColumn0, bigIntColumn1));
    ColumnMetadata oneLevelRow0 = new ColumnMetadata("onelevelrow0", oneLevelRowType);
    RowType twoLevelRowType = toRowType(ImmutableList.of(oneLevelRow0, bigIntColumn0, bigIntColumn1));
    ColumnMetadata twoLevelRow0 = new ColumnMetadata("twolevelrow0", twoLevelRowType);
    List<ColumnMetadata> columnsForApplyProjectionTest = ImmutableList.of(bigIntColumn0, bigIntColumn1, oneLevelRow0, twoLevelRow0);
    SchemaTableName tableName = temporaryTable("apply_projection_tester");
    doCreateEmptyTable(tableName, ORC, columnsForApplyProjectionTest);
    try (Transaction transaction = newTransaction()) {
        ConnectorSession session = newSession();
        ConnectorMetadata metadata = transaction.getMetadata();
        ConnectorTableHandle tableHandle = getTableHandle(metadata, tableName);
        List<ColumnHandle> columnHandles = metadata.getColumnHandles(session, tableHandle).values().stream().filter(columnHandle -> !((HiveColumnHandle) columnHandle).isHidden()).collect(toList());
        assertEquals(columnHandles.size(), columnsForApplyProjectionTest.size());
        Map<String, ColumnHandle> columnHandleMap = columnHandles.stream().collect(toImmutableMap(handle -> ((HiveColumnHandle) handle).getBaseColumnName(), Function.identity()));
        // Emulate symbols coming from the query plan and map them to column handles
        Map<String, ColumnHandle> columnHandlesWithSymbols = ImmutableMap.of("symbol_0", columnHandleMap.get("int0"), "symbol_1", columnHandleMap.get("int1"), "symbol_2", columnHandleMap.get("onelevelrow0"), "symbol_3", columnHandleMap.get("twolevelrow0"));
        // Create variables for the emulated symbols
        Map<String, Variable> symbolVariableMapping = columnHandlesWithSymbols.entrySet().stream().collect(toImmutableMap(Map.Entry::getKey, e -> new Variable(e.getKey(), ((HiveColumnHandle) e.getValue()).getBaseType())));
        // Create dereference expressions for testing
        FieldDereference symbol2Field0 = new FieldDereference(BIGINT, symbolVariableMapping.get("symbol_2"), 0);
        FieldDereference symbol3Field0 = new FieldDereference(oneLevelRowType, symbolVariableMapping.get("symbol_3"), 0);
        FieldDereference symbol3Field0Field0 = new FieldDereference(BIGINT, symbol3Field0, 0);
        FieldDereference symbol3Field1 = new FieldDereference(BIGINT, symbolVariableMapping.get("symbol_3"), 1);
        Map<String, ColumnHandle> inputAssignments;
        List<ConnectorExpression> inputProjections;
        Optional<ProjectionApplicationResult<ConnectorTableHandle>> projectionResult;
        List<ConnectorExpression> expectedProjections;
        Map<String, Type> expectedAssignments;
        // Test projected columns pushdown to HiveTableHandle in case of all variable references
        inputAssignments = getColumnHandlesFor(columnHandlesWithSymbols, ImmutableList.of("symbol_0", "symbol_1"));
        inputProjections = ImmutableList.of(symbolVariableMapping.get("symbol_0"), symbolVariableMapping.get("symbol_1"));
        expectedAssignments = ImmutableMap.of("symbol_0", BIGINT, "symbol_1", BIGINT);
        projectionResult = metadata.applyProjection(session, tableHandle, inputProjections, inputAssignments);
        assertProjectionResult(projectionResult, false, inputProjections, expectedAssignments);
        // Empty result when projected column handles are same as those present in table handle
        projectionResult = metadata.applyProjection(session, projectionResult.get().getHandle(), inputProjections, inputAssignments);
        assertProjectionResult(projectionResult, true, ImmutableList.of(), ImmutableMap.of());
        // Extra columns handles in HiveTableHandle should get pruned
        projectionResult = metadata.applyProjection(session, ((HiveTableHandle) tableHandle).withProjectedColumns(ImmutableSet.copyOf(columnHandles)), inputProjections, inputAssignments);
        assertProjectionResult(projectionResult, false, inputProjections, expectedAssignments);
        // Test projection pushdown for dereferences
        inputAssignments = getColumnHandlesFor(columnHandlesWithSymbols, ImmutableList.of("symbol_2", "symbol_3"));
        inputProjections = ImmutableList.of(symbol2Field0, symbol3Field0Field0, symbol3Field1);
        expectedAssignments = ImmutableMap.of("onelevelrow0#f_int0", BIGINT, "twolevelrow0#f_onelevelrow0#f_int0", BIGINT, "twolevelrow0#f_int0", BIGINT);
        expectedProjections = ImmutableList.of(new Variable("onelevelrow0#f_int0", BIGINT), new Variable("twolevelrow0#f_onelevelrow0#f_int0", BIGINT), new Variable("twolevelrow0#f_int0", BIGINT));
        projectionResult = metadata.applyProjection(session, tableHandle, inputProjections, inputAssignments);
        assertProjectionResult(projectionResult, false, expectedProjections, expectedAssignments);
        // Test reuse of virtual column handles
        // Round-1: input projections [symbol_2, symbol_2.int0]. virtual handle is created for symbol_2.int0.
        inputAssignments = getColumnHandlesFor(columnHandlesWithSymbols, ImmutableList.of("symbol_2"));
        inputProjections = ImmutableList.of(symbol2Field0, symbolVariableMapping.get("symbol_2"));
        projectionResult = metadata.applyProjection(session, tableHandle, inputProjections, inputAssignments);
        expectedProjections = ImmutableList.of(new Variable("onelevelrow0#f_int0", BIGINT), symbolVariableMapping.get("symbol_2"));
        expectedAssignments = ImmutableMap.of("onelevelrow0#f_int0", BIGINT, "symbol_2", oneLevelRowType);
        assertProjectionResult(projectionResult, false, expectedProjections, expectedAssignments);
        // Round-2: input projections [symbol_2.int0 and onelevelrow0#f_int0]. Virtual handle is reused.
        Assignment newlyCreatedColumn = getOnlyElement(projectionResult.get().getAssignments().stream().filter(handle -> handle.getVariable().equals("onelevelrow0#f_int0")).collect(toList()));
        inputAssignments = ImmutableMap.<String, ColumnHandle>builder().putAll(getColumnHandlesFor(columnHandlesWithSymbols, ImmutableList.of("symbol_2"))).put(newlyCreatedColumn.getVariable(), newlyCreatedColumn.getColumn()).buildOrThrow();
        inputProjections = ImmutableList.of(symbol2Field0, new Variable("onelevelrow0#f_int0", BIGINT));
        projectionResult = metadata.applyProjection(session, tableHandle, inputProjections, inputAssignments);
        expectedProjections = ImmutableList.of(new Variable("onelevelrow0#f_int0", BIGINT), new Variable("onelevelrow0#f_int0", BIGINT));
        expectedAssignments = ImmutableMap.of("onelevelrow0#f_int0", BIGINT);
        assertProjectionResult(projectionResult, false, expectedProjections, expectedAssignments);
    } finally {
        dropTable(tableName);
    }
}
Also used : Assertions.assertInstanceOf(io.airlift.testing.Assertions.assertInstanceOf) FileSystem(org.apache.hadoop.fs.FileSystem) Test(org.testng.annotations.Test) FileStatus(org.apache.hadoop.fs.FileStatus) NOT_SUPPORTED(io.trino.spi.StandardErrorCode.NOT_SUPPORTED) TableNotFoundException(io.trino.spi.connector.TableNotFoundException) HiveColumnStatistics.createDateColumnStatistics(io.trino.plugin.hive.metastore.HiveColumnStatistics.createDateColumnStatistics) Files.createTempDirectory(java.nio.file.Files.createTempDirectory) Map(java.util.Map) PRESTO_QUERY_ID_NAME(io.trino.plugin.hive.HiveMetadata.PRESTO_QUERY_ID_NAME) ConnectorPageSource(io.trino.spi.connector.ConnectorPageSource) ViewNotFoundException(io.trino.spi.connector.ViewNotFoundException) MaterializedRow(io.trino.testing.MaterializedRow) ROLLBACK_AFTER_FINISH_INSERT(io.trino.plugin.hive.AbstractTestHive.TransactionDeleteInsertTestTag.ROLLBACK_AFTER_FINISH_INSERT) ENGLISH(java.util.Locale.ENGLISH) Assert.assertFalse(org.testng.Assert.assertFalse) HiveIdentity(io.trino.plugin.hive.authentication.HiveIdentity) Domain(io.trino.spi.predicate.Domain) MANAGED_TABLE(org.apache.hadoop.hive.metastore.TableType.MANAGED_TABLE) ASCENDING(io.trino.plugin.hive.metastore.SortingColumn.Order.ASCENDING) ValueSet(io.trino.spi.predicate.ValueSet) MoreExecutors.directExecutor(com.google.common.util.concurrent.MoreExecutors.directExecutor) COMMIT(io.trino.plugin.hive.AbstractTestHive.TransactionDeleteInsertTestTag.COMMIT) NOT_PARTITIONED(io.trino.spi.connector.NotPartitionedPartitionHandle.NOT_PARTITIONED) Lists.newArrayList(com.google.common.collect.Lists.newArrayList) DESCENDING(io.trino.plugin.hive.metastore.SortingColumn.Order.DESCENDING) ConnectorPartitioningHandle(io.trino.spi.connector.ConnectorPartitioningHandle) TrinoS3ConfigurationInitializer(io.trino.plugin.hive.s3.TrinoS3ConfigurationInitializer) CatalogSchemaTableName(io.trino.spi.connector.CatalogSchemaTableName) MetastoreLocator(io.trino.plugin.hive.metastore.thrift.MetastoreLocator) ROLLBACK_AFTER_SINK_FINISH(io.trino.plugin.hive.AbstractTestHive.TransactionDeleteInsertTestTag.ROLLBACK_AFTER_SINK_FINISH) TableScanRedirectApplicationResult(io.trino.spi.connector.TableScanRedirectApplicationResult) TableColumnsMetadata(io.trino.spi.connector.TableColumnsMetadata) REAL(io.trino.spi.type.RealType.REAL) Partition(io.trino.plugin.hive.metastore.Partition) BUCKETED_BY_PROPERTY(io.trino.plugin.hive.HiveTableProperties.BUCKETED_BY_PROPERTY) ColumnMetadata(io.trino.spi.connector.ColumnMetadata) TIMESTAMP_MILLIS(io.trino.spi.type.TimestampType.TIMESTAMP_MILLIS) LocalDateTime(java.time.LocalDateTime) ConnectorTableMetadata(io.trino.spi.connector.ConnectorTableMetadata) HiveBasicStatistics.createEmptyStatistics(io.trino.plugin.hive.HiveBasicStatistics.createEmptyStatistics) Variable(io.trino.spi.expression.Variable) StorageFormat.fromHiveStorageFormat(io.trino.plugin.hive.metastore.StorageFormat.fromHiveStorageFormat) OptionalLong(java.util.OptionalLong) VARCHAR(io.trino.spi.type.VarcharType.VARCHAR) HiveMetastore(io.trino.plugin.hive.metastore.HiveMetastore) OrcPageSource(io.trino.plugin.hive.orc.OrcPageSource) SEQUENCEFILE(io.trino.plugin.hive.HiveStorageFormat.SEQUENCEFILE) ScheduledExecutorService(java.util.concurrent.ScheduledExecutorService) HIVE_INVALID_PARTITION_VALUE(io.trino.plugin.hive.HiveErrorCode.HIVE_INVALID_PARTITION_VALUE) ImmutableSet.toImmutableSet(com.google.common.collect.ImmutableSet.toImmutableSet) Assertions.assertGreaterThanOrEqual(io.airlift.testing.Assertions.assertGreaterThanOrEqual) ImmutableMultimap(com.google.common.collect.ImmutableMultimap) HiveSessionProperties.isTemporaryStagingDirectoryEnabled(io.trino.plugin.hive.HiveSessionProperties.isTemporaryStagingDirectoryEnabled) AfterClass(org.testng.annotations.AfterClass) HiveAzureConfig(io.trino.plugin.hive.azure.HiveAzureConfig) FileUtils.makePartName(org.apache.hadoop.hive.common.FileUtils.makePartName) MapType(io.trino.spi.type.MapType) ConnectorSplit(io.trino.spi.connector.ConnectorSplit) TRANSACTIONAL(io.trino.plugin.hive.HiveTableProperties.TRANSACTIONAL) SPARK_TABLE_PROVIDER_KEY(io.trino.plugin.hive.util.HiveUtil.SPARK_TABLE_PROVIDER_KEY) ConnectorSplitSource(io.trino.spi.connector.ConnectorSplitSource) IOException(java.io.IOException) Iterables.getOnlyElement(com.google.common.collect.Iterables.getOnlyElement) STAGE_AND_MOVE_TO_TARGET_DIRECTORY(io.trino.plugin.hive.LocationHandle.WriteMode.STAGE_AND_MOVE_TO_TARGET_DIRECTORY) ROLLBACK_AFTER_BEGIN_INSERT(io.trino.plugin.hive.AbstractTestHive.TransactionDeleteInsertTestTag.ROLLBACK_AFTER_BEGIN_INSERT) HostAndPort(com.google.common.net.HostAndPort) CatalogName(io.trino.plugin.base.CatalogName) DOUBLE(io.trino.spi.type.DoubleType.DOUBLE) HIVE_INT(io.trino.plugin.hive.HiveType.HIVE_INT) ConnectorTableProperties(io.trino.spi.connector.ConnectorTableProperties) ConnectorExpression(io.trino.spi.expression.ConnectorExpression) HiveTestUtils.mapType(io.trino.plugin.hive.HiveTestUtils.mapType) ParquetPageSource(io.trino.plugin.hive.parquet.ParquetPageSource) ThriftMetastoreConfig(io.trino.plugin.hive.metastore.thrift.ThriftMetastoreConfig) RCTEXT(io.trino.plugin.hive.HiveStorageFormat.RCTEXT) VarcharType.createVarcharType(io.trino.spi.type.VarcharType.createVarcharType) WriteInfo(io.trino.plugin.hive.LocationService.WriteInfo) HivePrivilege(io.trino.plugin.hive.metastore.HivePrivilegeInfo.HivePrivilege) PARTITION_KEY(io.trino.plugin.hive.HiveColumnHandle.ColumnType.PARTITION_KEY) HiveColumnStatistics.createStringColumnStatistics(io.trino.plugin.hive.metastore.HiveColumnStatistics.createStringColumnStatistics) MoreFiles.deleteRecursively(com.google.common.io.MoreFiles.deleteRecursively) MaterializedResult(io.trino.testing.MaterializedResult) HiveUtil.toPartitionValues(io.trino.plugin.hive.util.HiveUtil.toPartitionValues) NO_RETRIES(io.trino.spi.connector.RetryMode.NO_RETRIES) ConnectorMaterializedViewDefinition(io.trino.spi.connector.ConnectorMaterializedViewDefinition) ICEBERG_TABLE_TYPE_VALUE(io.trino.plugin.hive.util.HiveUtil.ICEBERG_TABLE_TYPE_VALUE) Duration(io.airlift.units.Duration) BUCKETING_V1(io.trino.plugin.hive.util.HiveBucketing.BucketingVersion.BUCKETING_V1) SqlTimestamp(io.trino.spi.type.SqlTimestamp) ICEBERG_TABLE_TYPE_NAME(io.trino.plugin.hive.util.HiveUtil.ICEBERG_TABLE_TYPE_NAME) NOOP_METADATA_PROVIDER(io.trino.spi.connector.MetadataProvider.NOOP_METADATA_PROVIDER) HiveMetastoreFactory(io.trino.plugin.hive.metastore.HiveMetastoreFactory) Preconditions.checkArgument(com.google.common.base.Preconditions.checkArgument) HiveTestUtils.arrayType(io.trino.plugin.hive.HiveTestUtils.arrayType) Block(io.trino.spi.block.Block) HIVE_PARTITION_SCHEMA_MISMATCH(io.trino.plugin.hive.HiveErrorCode.HIVE_PARTITION_SCHEMA_MISMATCH) ConnectorViewDefinition(io.trino.spi.connector.ConnectorViewDefinition) INTEGER(io.trino.spi.type.IntegerType.INTEGER) ROLLBACK_RIGHT_AWAY(io.trino.plugin.hive.AbstractTestHive.TransactionDeleteInsertTestTag.ROLLBACK_RIGHT_AWAY) ImmutableSet(com.google.common.collect.ImmutableSet) BeforeClass(org.testng.annotations.BeforeClass) Collection(java.util.Collection) UUID(java.util.UUID) Assert.assertNotNull(org.testng.Assert.assertNotNull) TrinoAzureConfigurationInitializer(io.trino.plugin.hive.azure.TrinoAzureConfigurationInitializer) HiveColumnStatistics.createDoubleColumnStatistics(io.trino.plugin.hive.metastore.HiveColumnStatistics.createDoubleColumnStatistics) BUCKET_COLUMN_NAME(io.trino.plugin.hive.HiveColumnHandle.BUCKET_COLUMN_NAME) BIGINT(io.trino.spi.type.BigintType.BIGINT) SORTED_BY_PROPERTY(io.trino.plugin.hive.HiveTableProperties.SORTED_BY_PROPERTY) LocalDate(java.time.LocalDate) JsonCodec(io.airlift.json.JsonCodec) IntStream(java.util.stream.IntStream) HiveTestUtils.getDefaultHivePageSourceFactories(io.trino.plugin.hive.HiveTestUtils.getDefaultHivePageSourceFactories) Constraint(io.trino.spi.connector.Constraint) Assert.assertNull(org.testng.Assert.assertNull) OptionalDouble(java.util.OptionalDouble) Assert.assertEquals(org.testng.Assert.assertEquals) OptionalInt(java.util.OptionalInt) Function(java.util.function.Function) HashSet(java.util.HashSet) HYPER_LOG_LOG(io.trino.spi.type.HyperLogLogType.HYPER_LOG_LOG) ImmutableList(com.google.common.collect.ImmutableList) BridgingHiveMetastore(io.trino.plugin.hive.metastore.thrift.BridgingHiveMetastore) TableStatistics(io.trino.spi.statistics.TableStatistics) HiveColumnHandle.createBaseColumn(io.trino.plugin.hive.HiveColumnHandle.createBaseColumn) Math.toIntExact(java.lang.Math.toIntExact) ExecutorService(java.util.concurrent.ExecutorService) ConnectorPageSink(io.trino.spi.connector.ConnectorPageSink) DESC_NULLS_LAST(io.trino.spi.connector.SortOrder.DESC_NULLS_LAST) ORC(io.trino.plugin.hive.HiveStorageFormat.ORC) HiveColumnStatistics.createIntegerColumnStatistics(io.trino.plugin.hive.metastore.HiveColumnStatistics.createIntegerColumnStatistics) UTF_8(java.nio.charset.StandardCharsets.UTF_8) Assert.fail(org.testng.Assert.fail) ConnectorPageSourceProvider(io.trino.spi.connector.ConnectorPageSourceProvider) DateTime(org.joda.time.DateTime) PAGE_SORTER(io.trino.plugin.hive.HiveTestUtils.PAGE_SORTER) Executors.newFixedThreadPool(java.util.concurrent.Executors.newFixedThreadPool) HIVE_STRING(io.trino.plugin.hive.HiveType.HIVE_STRING) Hashing.sha256(com.google.common.hash.Hashing.sha256) HiveTestUtils.getHiveSession(io.trino.plugin.hive.HiveTestUtils.getHiveSession) Collectors.toList(java.util.stream.Collectors.toList) PRESTO_VERSION_NAME(io.trino.plugin.hive.HiveMetadata.PRESTO_VERSION_NAME) Assert.assertTrue(org.testng.Assert.assertTrue) PrincipalPrivileges(io.trino.plugin.hive.metastore.PrincipalPrivileges) ConnectorPageSinkProvider(io.trino.spi.connector.ConnectorPageSinkProvider) ConnectorTransactionHandle(io.trino.spi.connector.ConnectorTransactionHandle) ConnectorSplitManager(io.trino.spi.connector.ConnectorSplitManager) Arrays(java.util.Arrays) NamedTypeSignature(io.trino.spi.type.NamedTypeSignature) USER(io.trino.spi.security.PrincipalType.USER) Maps.uniqueIndex(com.google.common.collect.Maps.uniqueIndex) NO_ACID_TRANSACTION(io.trino.plugin.hive.acid.AcidTransaction.NO_ACID_TRANSACTION) HiveColumnStatistics.createDecimalColumnStatistics(io.trino.plugin.hive.metastore.HiveColumnStatistics.createDecimalColumnStatistics) TypeOperators(io.trino.spi.type.TypeOperators) ROLLBACK_AFTER_APPEND_PAGE(io.trino.plugin.hive.AbstractTestHive.TransactionDeleteInsertTestTag.ROLLBACK_AFTER_APPEND_PAGE) HiveTestUtils.rowType(io.trino.plugin.hive.HiveTestUtils.rowType) TrinoExceptionAssert.assertTrinoExceptionThrownBy(io.trino.testing.assertions.TrinoExceptionAssert.assertTrinoExceptionThrownBy) CharType.createCharType(io.trino.spi.type.CharType.createCharType) BigDecimal(java.math.BigDecimal) TypeId(io.trino.spi.type.TypeId) Sets.difference(com.google.common.collect.Sets.difference) PARQUET(io.trino.plugin.hive.HiveStorageFormat.PARQUET) Column(io.trino.plugin.hive.metastore.Column) Executors.newScheduledThreadPool(java.util.concurrent.Executors.newScheduledThreadPool) ConnectorOutputTableHandle(io.trino.spi.connector.ConnectorOutputTableHandle) ConnectorTableHandle(io.trino.spi.connector.ConnectorTableHandle) ViewColumn(io.trino.spi.connector.ConnectorViewDefinition.ViewColumn) ProjectionApplicationResult(io.trino.spi.connector.ProjectionApplicationResult) Slices.utf8Slice(io.airlift.slice.Slices.utf8Slice) PartitionWithStatistics(io.trino.plugin.hive.metastore.PartitionWithStatistics) SMALLINT(io.trino.spi.type.SmallintType.SMALLINT) HiveTestUtils.getDefaultHiveRecordCursorProviders(io.trino.plugin.hive.HiveTestUtils.getDefaultHiveRecordCursorProviders) ConnectorNodePartitioningProvider(io.trino.spi.connector.ConnectorNodePartitioningProvider) Table(io.trino.plugin.hive.metastore.Table) TestingNodeManager(io.trino.testing.TestingNodeManager) Range(io.trino.spi.predicate.Range) ImmutableList.toImmutableList(com.google.common.collect.ImmutableList.toImmutableList) PARTITIONED_BY_PROPERTY(io.trino.plugin.hive.HiveTableProperties.PARTITIONED_BY_PROPERTY) RcFilePageSource(io.trino.plugin.hive.rcfile.RcFilePageSource) Set(java.util.Set) MILLISECONDS(java.util.concurrent.TimeUnit.MILLISECONDS) SchemaTableName(io.trino.spi.connector.SchemaTableName) SortingProperty(io.trino.spi.connector.SortingProperty) ImmutableMap.toImmutableMap(com.google.common.collect.ImmutableMap.toImmutableMap) HiveColumnStatistics.createBooleanColumnStatistics(io.trino.plugin.hive.metastore.HiveColumnStatistics.createBooleanColumnStatistics) SchemaTablePrefix(io.trino.spi.connector.SchemaTablePrefix) Lists.reverse(com.google.common.collect.Lists.reverse) DATE(io.trino.spi.type.DateType.DATE) MoreObjects.toStringHelper(com.google.common.base.MoreObjects.toStringHelper) HiveColumnHandle.bucketColumnHandle(io.trino.plugin.hive.HiveColumnHandle.bucketColumnHandle) HivePrincipal(io.trino.plugin.hive.metastore.HivePrincipal) ConnectorTableLayout(io.trino.spi.connector.ConnectorTableLayout) ConnectorInsertTableHandle(io.trino.spi.connector.ConnectorInsertTableHandle) Slice(io.airlift.slice.Slice) NullableValue(io.trino.spi.predicate.NullableValue) Page(io.trino.spi.Page) BOOLEAN(io.trino.spi.type.BooleanType.BOOLEAN) MINUTES(java.util.concurrent.TimeUnit.MINUTES) JoinCompiler(io.trino.sql.gen.JoinCompiler) HiveUtil.columnExtraInfo(io.trino.plugin.hive.util.HiveUtil.columnExtraInfo) GroupByHashPageIndexerFactory(io.trino.operator.GroupByHashPageIndexerFactory) Float.floatToRawIntBits(java.lang.Float.floatToRawIntBits) ALLOW_INSECURE(com.google.common.io.RecursiveDeleteOption.ALLOW_INSECURE) UNGROUPED_SCHEDULING(io.trino.spi.connector.ConnectorSplitManager.SplitSchedulingStrategy.UNGROUPED_SCHEDULING) ThreadLocalRandom(java.util.concurrent.ThreadLocalRandom) ColumnHandle(io.trino.spi.connector.ColumnHandle) TEXTFILE(io.trino.plugin.hive.HiveStorageFormat.TEXTFILE) VARBINARY(io.trino.spi.type.VarbinaryType.VARBINARY) HIVE_INVALID_BUCKET_FILES(io.trino.plugin.hive.HiveErrorCode.HIVE_INVALID_BUCKET_FILES) HiveType.toHiveType(io.trino.plugin.hive.HiveType.toHiveType) TestingMetastoreLocator(io.trino.plugin.hive.metastore.thrift.TestingMetastoreLocator) STORAGE_FORMAT_PROPERTY(io.trino.plugin.hive.HiveTableProperties.STORAGE_FORMAT_PROPERTY) GoogleGcsConfigurationInitializer(io.trino.plugin.hive.gcs.GoogleGcsConfigurationInitializer) ConstraintApplicationResult(io.trino.spi.connector.ConstraintApplicationResult) RecordCursor(io.trino.spi.connector.RecordCursor) BlockTypeOperators(io.trino.type.BlockTypeOperators) LongStream(java.util.stream.LongStream) NO_REDIRECTIONS(io.trino.plugin.hive.HiveTableRedirectionsProvider.NO_REDIRECTIONS) DecimalType.createDecimalType(io.trino.spi.type.DecimalType.createDecimalType) HIVE_LONG(io.trino.plugin.hive.HiveType.HIVE_LONG) HiveTestUtils.getTypes(io.trino.plugin.hive.HiveTestUtils.getTypes) TRANSACTION_CONFLICT(io.trino.spi.StandardErrorCode.TRANSACTION_CONFLICT) TESTING_TYPE_MANAGER(io.trino.type.InternalTypeManager.TESTING_TYPE_MANAGER) ConnectorSession(io.trino.spi.connector.ConnectorSession) MoreFutures.getFutureValue(io.airlift.concurrent.MoreFutures.getFutureValue) UTC(org.joda.time.DateTimeZone.UTC) SESSION(io.trino.plugin.hive.HiveTestUtils.SESSION) SqlVarbinary(io.trino.spi.type.SqlVarbinary) DiscretePredicates(io.trino.spi.connector.DiscretePredicates) CharType(io.trino.spi.type.CharType) TableType(org.apache.hadoop.hive.metastore.TableType) CachingHiveMetastore.cachingHiveMetastore(io.trino.plugin.hive.metastore.cache.CachingHiveMetastore.cachingHiveMetastore) TINYINT(io.trino.spi.type.TinyintType.TINYINT) DateTimeTestingUtils.sqlTimestampOf(io.trino.testing.DateTimeTestingUtils.sqlTimestampOf) Assertions.assertThat(org.assertj.core.api.Assertions.assertThat) Assertions.assertGreaterThan(io.airlift.testing.Assertions.assertGreaterThan) HiveColumnStatistics.createBinaryColumnStatistics(io.trino.plugin.hive.metastore.HiveColumnStatistics.createBinaryColumnStatistics) MoreCollectors.onlyElement(com.google.common.collect.MoreCollectors.onlyElement) NoHdfsAuthentication(io.trino.plugin.hive.authentication.NoHdfsAuthentication) QueryAssertions.assertEqualsIgnoreOrder(io.trino.testing.QueryAssertions.assertEqualsIgnoreOrder) HiveGcsConfig(io.trino.plugin.hive.gcs.HiveGcsConfig) Iterables.concat(com.google.common.collect.Iterables.concat) Path(org.apache.hadoop.fs.Path) KILOBYTE(io.airlift.units.DataSize.Unit.KILOBYTE) TIMESTAMP_WITH_TIME_ZONE(io.trino.spi.type.TimestampWithTimeZoneType.TIMESTAMP_WITH_TIME_ZONE) AVRO(io.trino.plugin.hive.HiveStorageFormat.AVRO) StorageFormat(io.trino.plugin.hive.metastore.StorageFormat) RowType(io.trino.spi.type.RowType) HiveWriteUtils.getTableDefaultLocation(io.trino.plugin.hive.util.HiveWriteUtils.getTableDefaultLocation) ImmutableMap(com.google.common.collect.ImmutableMap) Predicate(java.util.function.Predicate) ThriftHiveMetastore(io.trino.plugin.hive.metastore.thrift.ThriftHiveMetastore) HiveSessionProperties.getTemporaryStagingDirectoryPath(io.trino.plugin.hive.HiveSessionProperties.getTemporaryStagingDirectoryPath) TrinoException(io.trino.spi.TrinoException) ArrayType(io.trino.spi.type.ArrayType) MaterializedResult.materializeSourceDataStream(io.trino.testing.MaterializedResult.materializeSourceDataStream) ConnectorBucketNodeMap(io.trino.spi.connector.ConnectorBucketNodeMap) ASC_NULLS_FIRST(io.trino.spi.connector.SortOrder.ASC_NULLS_FIRST) String.format(java.lang.String.format) SqlDate(io.trino.spi.type.SqlDate) Preconditions.checkState(com.google.common.base.Preconditions.checkState) SqlTimestampWithTimeZone(io.trino.spi.type.SqlTimestampWithTimeZone) DataSize(io.airlift.units.DataSize) HdfsContext(io.trino.plugin.hive.HdfsEnvironment.HdfsContext) List(java.util.List) DynamicFilter(io.trino.spi.connector.DynamicFilter) Assignment(io.trino.spi.connector.Assignment) Optional(java.util.Optional) ConnectorMetadata(io.trino.spi.connector.ConnectorMetadata) HivePrivilegeInfo(io.trino.plugin.hive.metastore.HivePrivilegeInfo) SqlStandardAccessControlMetadata(io.trino.plugin.hive.security.SqlStandardAccessControlMetadata) Logger(io.airlift.log.Logger) MetastoreConfig(io.trino.plugin.hive.metastore.MetastoreConfig) Type(io.trino.spi.type.Type) VarcharType.createUnboundedVarcharType(io.trino.spi.type.VarcharType.createUnboundedVarcharType) CounterStat(io.airlift.stats.CounterStat) HashMap(java.util.HashMap) HiveBasicStatistics.createZeroStatistics(io.trino.plugin.hive.HiveBasicStatistics.createZeroStatistics) CSV(io.trino.plugin.hive.HiveStorageFormat.CSV) AtomicReference(java.util.concurrent.atomic.AtomicReference) VarcharType(io.trino.spi.type.VarcharType) HiveColumnStatistics(io.trino.plugin.hive.metastore.HiveColumnStatistics) ROLLBACK_AFTER_DELETE(io.trino.plugin.hive.AbstractTestHive.TransactionDeleteInsertTestTag.ROLLBACK_AFTER_DELETE) Assertions.assertThatThrownBy(org.assertj.core.api.Assertions.assertThatThrownBy) Verify.verify(com.google.common.base.Verify.verify) Assertions.assertLessThanOrEqual(io.airlift.testing.Assertions.assertLessThanOrEqual) Threads.daemonThreadsNamed(io.airlift.concurrent.Threads.daemonThreadsNamed) SemiTransactionalHiveMetastore(io.trino.plugin.hive.metastore.SemiTransactionalHiveMetastore) RecordPageSource(io.trino.spi.connector.RecordPageSource) Objects.requireNonNull(java.util.Objects.requireNonNull) RowFieldName(io.trino.spi.type.RowFieldName) JSON(io.trino.plugin.hive.HiveStorageFormat.JSON) RCBINARY(io.trino.plugin.hive.HiveStorageFormat.RCBINARY) NO_PRIVILEGES(io.trino.plugin.hive.metastore.PrincipalPrivileges.NO_PRIVILEGES) DELTA_LAKE_PROVIDER(io.trino.plugin.hive.util.HiveUtil.DELTA_LAKE_PROVIDER) ColumnStatistics(io.trino.spi.statistics.ColumnStatistics) FieldDereference(io.trino.spi.expression.FieldDereference) HiveTestUtils.getDefaultHiveFileWriterFactories(io.trino.plugin.hive.HiveTestUtils.getDefaultHiveFileWriterFactories) HiveTestUtils.getHiveSessionProperties(io.trino.plugin.hive.HiveTestUtils.getHiveSessionProperties) TupleDomain(io.trino.spi.predicate.TupleDomain) HiveWriteUtils.createDirectory(io.trino.plugin.hive.util.HiveWriteUtils.createDirectory) TestingConnectorSession(io.trino.testing.TestingConnectorSession) HiveS3Config(io.trino.plugin.hive.s3.HiveS3Config) Executors.newCachedThreadPool(java.util.concurrent.Executors.newCachedThreadPool) BUCKET_COUNT_PROPERTY(io.trino.plugin.hive.HiveTableProperties.BUCKET_COUNT_PROPERTY) SortingColumn(io.trino.plugin.hive.metastore.SortingColumn) SECONDS(java.util.concurrent.TimeUnit.SECONDS) REGULAR(io.trino.plugin.hive.HiveColumnHandle.ColumnType.REGULAR) ColumnMetadata(io.trino.spi.connector.ColumnMetadata) Variable(io.trino.spi.expression.Variable) ConnectorExpression(io.trino.spi.expression.ConnectorExpression) RowType(io.trino.spi.type.RowType) Assignment(io.trino.spi.connector.Assignment) FieldDereference(io.trino.spi.expression.FieldDereference) ConnectorSession(io.trino.spi.connector.ConnectorSession) TestingConnectorSession(io.trino.testing.TestingConnectorSession) ConnectorMetadata(io.trino.spi.connector.ConnectorMetadata) HiveColumnHandle.bucketColumnHandle(io.trino.plugin.hive.HiveColumnHandle.bucketColumnHandle) ColumnHandle(io.trino.spi.connector.ColumnHandle) CatalogSchemaTableName(io.trino.spi.connector.CatalogSchemaTableName) SchemaTableName(io.trino.spi.connector.SchemaTableName) ProjectionApplicationResult(io.trino.spi.connector.ProjectionApplicationResult) ConnectorTableHandle(io.trino.spi.connector.ConnectorTableHandle) MapType(io.trino.spi.type.MapType) HiveTestUtils.mapType(io.trino.plugin.hive.HiveTestUtils.mapType) VarcharType.createVarcharType(io.trino.spi.type.VarcharType.createVarcharType) HiveTestUtils.arrayType(io.trino.plugin.hive.HiveTestUtils.arrayType) HiveTestUtils.rowType(io.trino.plugin.hive.HiveTestUtils.rowType) CharType.createCharType(io.trino.spi.type.CharType.createCharType) HiveType.toHiveType(io.trino.plugin.hive.HiveType.toHiveType) DecimalType.createDecimalType(io.trino.spi.type.DecimalType.createDecimalType) CharType(io.trino.spi.type.CharType) TableType(org.apache.hadoop.hive.metastore.TableType) RowType(io.trino.spi.type.RowType) ArrayType(io.trino.spi.type.ArrayType) Type(io.trino.spi.type.Type) VarcharType.createUnboundedVarcharType(io.trino.spi.type.VarcharType.createUnboundedVarcharType) VarcharType(io.trino.spi.type.VarcharType) Map(java.util.Map) ImmutableMap.toImmutableMap(com.google.common.collect.ImmutableMap.toImmutableMap) ImmutableMap(com.google.common.collect.ImmutableMap) ConnectorBucketNodeMap(io.trino.spi.connector.ConnectorBucketNodeMap) HashMap(java.util.HashMap) Test(org.testng.annotations.Test)

Example 3 with ProjectionApplicationResult

use of io.trino.spi.connector.ProjectionApplicationResult in project trino by trinodb.

the class TestPruneTableScanColumns method mockApplyProjection.

private Optional<ProjectionApplicationResult<ConnectorTableHandle>> mockApplyProjection(ConnectorSession session, ConnectorTableHandle tableHandle, List<ConnectorExpression> projections, Map<String, ColumnHandle> assignments) {
    MockConnectorTableHandle handle = (MockConnectorTableHandle) tableHandle;
    List<Variable> variables = projections.stream().map(Variable.class::cast).collect(toImmutableList());
    List<ColumnHandle> newColumns = variables.stream().map(variable -> assignments.get(variable.getName())).collect(toImmutableList());
    if (handle.getColumns().isPresent() && newColumns.equals(handle.getColumns().get())) {
        return Optional.empty();
    }
    return Optional.of(new ProjectionApplicationResult<>(new MockConnectorTableHandle(handle.getTableName(), handle.getConstraint(), Optional.of(newColumns)), projections, variables.stream().map(variable -> new Assignment(variable.getName(), assignments.get(variable.getName()), ((MockConnectorColumnHandle) assignments.get(variable.getName())).getType())).collect(toImmutableList()), false));
}
Also used : ColumnMetadata(io.trino.spi.connector.ColumnMetadata) PlanMatchPattern(io.trino.sql.planner.assertions.PlanMatchPattern) Test(org.testng.annotations.Test) Variable(io.trino.spi.expression.Variable) CatalogName(io.trino.connector.CatalogName) ImmutableList(com.google.common.collect.ImmutableList) MockConnectorFactory(io.trino.connector.MockConnectorFactory) TpchTableHandle(io.trino.plugin.tpch.TpchTableHandle) ConnectorTableHandle(io.trino.spi.connector.ConnectorTableHandle) Map(java.util.Map) ColumnHandle(io.trino.spi.connector.ColumnHandle) ProjectionApplicationResult(io.trino.spi.connector.ProjectionApplicationResult) Symbol(io.trino.sql.planner.Symbol) TINY_SCHEMA_NAME(io.trino.plugin.tpch.TpchMetadata.TINY_SCHEMA_NAME) TpchColumnHandle(io.trino.plugin.tpch.TpchColumnHandle) RuleTester.defaultRuleTester(io.trino.sql.planner.iterative.rule.test.RuleTester.defaultRuleTester) ImmutableMap(com.google.common.collect.ImmutableMap) BaseRuleTest(io.trino.sql.planner.iterative.rule.test.BaseRuleTest) TINY_SCALE_FACTOR(io.trino.plugin.tpch.TpchMetadata.TINY_SCALE_FACTOR) MockConnectorColumnHandle(io.trino.connector.MockConnectorColumnHandle) Domain(io.trino.spi.predicate.Domain) ImmutableList.toImmutableList(com.google.common.collect.ImmutableList.toImmutableList) Assignments(io.trino.sql.planner.plan.Assignments) ConnectorSession(io.trino.spi.connector.ConnectorSession) RuleTester(io.trino.sql.planner.iterative.rule.test.RuleTester) TupleDomain(io.trino.spi.predicate.TupleDomain) SchemaTableName(io.trino.spi.connector.SchemaTableName) TpchTransactionHandle(io.trino.plugin.tpch.TpchTransactionHandle) DOUBLE(io.trino.spi.type.DoubleType.DOUBLE) MockConnectorTableHandle(io.trino.connector.MockConnectorTableHandle) MockConnectorTransactionHandle(io.trino.connector.MockConnectorTransactionHandle) List(java.util.List) TestingSession.testSessionBuilder(io.trino.testing.TestingSession.testSessionBuilder) TableHandle(io.trino.metadata.TableHandle) ConnectorExpression(io.trino.spi.expression.ConnectorExpression) PlanMatchPattern.strictProject(io.trino.sql.planner.assertions.PlanMatchPattern.strictProject) Assignment(io.trino.spi.connector.Assignment) PlanMatchPattern.strictConstrainedTableScan(io.trino.sql.planner.assertions.PlanMatchPattern.strictConstrainedTableScan) Optional(java.util.Optional) TestingColumnHandle(io.trino.testing.TestingMetadata.TestingColumnHandle) PlanBuilder.expression(io.trino.sql.planner.iterative.rule.test.PlanBuilder.expression) DATE(io.trino.spi.type.DateType.DATE) PlanMatchPattern.tableScan(io.trino.sql.planner.assertions.PlanMatchPattern.tableScan) PlanMatchPattern.strictTableScan(io.trino.sql.planner.assertions.PlanMatchPattern.strictTableScan) Assignment(io.trino.spi.connector.Assignment) ColumnHandle(io.trino.spi.connector.ColumnHandle) TpchColumnHandle(io.trino.plugin.tpch.TpchColumnHandle) MockConnectorColumnHandle(io.trino.connector.MockConnectorColumnHandle) TestingColumnHandle(io.trino.testing.TestingMetadata.TestingColumnHandle) Variable(io.trino.spi.expression.Variable) MockConnectorTableHandle(io.trino.connector.MockConnectorTableHandle)

Example 4 with ProjectionApplicationResult

use of io.trino.spi.connector.ProjectionApplicationResult in project trino by trinodb.

the class IcebergMetadata method applyProjection.

@Override
public Optional<ProjectionApplicationResult<ConnectorTableHandle>> applyProjection(ConnectorSession session, ConnectorTableHandle handle, List<ConnectorExpression> projections, Map<String, ColumnHandle> assignments) {
    if (!isProjectionPushdownEnabled(session)) {
        return Optional.empty();
    }
    // Create projected column representations for supported sub expressions. Simple column references and chain of
    // dereferences on a variable are supported right now.
    Set<ConnectorExpression> projectedExpressions = projections.stream().flatMap(expression -> extractSupportedProjectedColumns(expression).stream()).collect(toImmutableSet());
    Map<ConnectorExpression, ProjectedColumnRepresentation> columnProjections = projectedExpressions.stream().collect(toImmutableMap(Function.identity(), HiveApplyProjectionUtil::createProjectedColumnRepresentation));
    IcebergTableHandle icebergTableHandle = (IcebergTableHandle) handle;
    // all references are simple variables
    if (columnProjections.values().stream().allMatch(ProjectedColumnRepresentation::isVariable)) {
        Set<IcebergColumnHandle> projectedColumns = assignments.values().stream().map(IcebergColumnHandle.class::cast).collect(toImmutableSet());
        if (icebergTableHandle.getProjectedColumns().equals(projectedColumns)) {
            return Optional.empty();
        }
        List<Assignment> assignmentsList = assignments.entrySet().stream().map(assignment -> new Assignment(assignment.getKey(), assignment.getValue(), ((IcebergColumnHandle) assignment.getValue()).getType())).collect(toImmutableList());
        return Optional.of(new ProjectionApplicationResult<>(icebergTableHandle.withProjectedColumns(projectedColumns), projections, assignmentsList, false));
    }
    Map<String, Assignment> newAssignments = new HashMap<>();
    ImmutableMap.Builder<ConnectorExpression, Variable> newVariablesBuilder = ImmutableMap.builder();
    ImmutableSet.Builder<IcebergColumnHandle> projectedColumnsBuilder = ImmutableSet.builder();
    for (Map.Entry<ConnectorExpression, ProjectedColumnRepresentation> entry : columnProjections.entrySet()) {
        ConnectorExpression expression = entry.getKey();
        ProjectedColumnRepresentation projectedColumn = entry.getValue();
        IcebergColumnHandle baseColumnHandle = (IcebergColumnHandle) assignments.get(projectedColumn.getVariable().getName());
        IcebergColumnHandle projectedColumnHandle = createProjectedColumnHandle(baseColumnHandle, projectedColumn.getDereferenceIndices(), expression.getType());
        String projectedColumnName = projectedColumnHandle.getQualifiedName();
        Variable projectedColumnVariable = new Variable(projectedColumnName, expression.getType());
        Assignment newAssignment = new Assignment(projectedColumnName, projectedColumnHandle, expression.getType());
        newAssignments.putIfAbsent(projectedColumnName, newAssignment);
        newVariablesBuilder.put(expression, projectedColumnVariable);
        projectedColumnsBuilder.add(projectedColumnHandle);
    }
    // Modify projections to refer to new variables
    Map<ConnectorExpression, Variable> newVariables = newVariablesBuilder.buildOrThrow();
    List<ConnectorExpression> newProjections = projections.stream().map(expression -> replaceWithNewVariables(expression, newVariables)).collect(toImmutableList());
    List<Assignment> outputAssignments = newAssignments.values().stream().collect(toImmutableList());
    return Optional.of(new ProjectionApplicationResult<>(icebergTableHandle.withProjectedColumns(projectedColumnsBuilder.build()), newProjections, outputAssignments, false));
}
Also used : IcebergUtil.getPartitionKeys(io.trino.plugin.iceberg.IcebergUtil.getPartitionKeys) TrinoCatalog(io.trino.plugin.iceberg.catalog.TrinoCatalog) FileSystem(org.apache.hadoop.fs.FileSystem) ConnectorTableExecuteHandle(io.trino.spi.connector.ConnectorTableExecuteHandle) HiveApplyProjectionUtil.replaceWithNewVariables(io.trino.plugin.hive.HiveApplyProjectionUtil.replaceWithNewVariables) Collections.singletonList(java.util.Collections.singletonList) NOT_SUPPORTED(io.trino.spi.StandardErrorCode.NOT_SUPPORTED) TableNotFoundException(io.trino.spi.connector.TableNotFoundException) Matcher(java.util.regex.Matcher) ConnectorOutputTableHandle(io.trino.spi.connector.ConnectorOutputTableHandle) ConnectorTableHandle(io.trino.spi.connector.ConnectorTableHandle) Map(java.util.Map) RewriteFiles(org.apache.iceberg.RewriteFiles) ProjectionApplicationResult(io.trino.spi.connector.ProjectionApplicationResult) HdfsEnvironment(io.trino.plugin.hive.HdfsEnvironment) CloseableIterable(org.apache.iceberg.io.CloseableIterable) IcebergUtil.newCreateTableTransaction(io.trino.plugin.iceberg.IcebergUtil.newCreateTableTransaction) Domain(io.trino.spi.predicate.Domain) ImmutableList.toImmutableList(com.google.common.collect.ImmutableList.toImmutableList) IcebergTableExecuteHandle(io.trino.plugin.iceberg.procedure.IcebergTableExecuteHandle) Set(java.util.Set) Schema(org.apache.iceberg.Schema) ColumnIdentity.primitiveColumnIdentity(io.trino.plugin.iceberg.ColumnIdentity.primitiveColumnIdentity) SchemaTableName(io.trino.spi.connector.SchemaTableName) PartitionSpecParser(org.apache.iceberg.PartitionSpecParser) Collectors.joining(java.util.stream.Collectors.joining) Type(org.apache.iceberg.types.Type) UncheckedIOException(java.io.UncheckedIOException) ImmutableMap.toImmutableMap(com.google.common.collect.ImmutableMap.toImmutableMap) TrinoPrincipal(io.trino.spi.security.TrinoPrincipal) CatalogSchemaTableName(io.trino.spi.connector.CatalogSchemaTableName) SchemaTablePrefix(io.trino.spi.connector.SchemaTablePrefix) RemoteIterator(org.apache.hadoop.fs.RemoteIterator) Iterables(com.google.common.collect.Iterables) ConnectorTableLayout(io.trino.spi.connector.ConnectorTableLayout) ConnectorInsertTableHandle(io.trino.spi.connector.ConnectorInsertTableHandle) IcebergUtil.deserializePartitionValue(io.trino.plugin.iceberg.IcebergUtil.deserializePartitionValue) Slice(io.airlift.slice.Slice) NullableValue(io.trino.spi.predicate.NullableValue) ColumnMetadata(io.trino.spi.connector.ColumnMetadata) ConnectorTableMetadata(io.trino.spi.connector.ConnectorTableMetadata) Variable(io.trino.spi.expression.Variable) Supplier(java.util.function.Supplier) OptionalLong(java.util.OptionalLong) MaterializedViewFreshness(io.trino.spi.connector.MaterializedViewFreshness) ColumnHandle(io.trino.spi.connector.ColumnHandle) ImmutableSet.toImmutableSet(com.google.common.collect.ImmutableSet.toImmutableSet) FILE_FORMAT_PROPERTY(io.trino.plugin.iceberg.IcebergTableProperties.FILE_FORMAT_PROPERTY) OPTIMIZE(io.trino.plugin.iceberg.procedure.IcebergTableProcedureId.OPTIMIZE) ConstraintApplicationResult(io.trino.spi.connector.ConstraintApplicationResult) DEPENDS_ON_TABLES(io.trino.plugin.iceberg.catalog.hms.TrinoHiveCatalog.DEPENDS_ON_TABLES) Table(org.apache.iceberg.Table) IOException(java.io.IOException) ConnectorSession(io.trino.spi.connector.ConnectorSession) ICEBERG_INVALID_METADATA(io.trino.plugin.iceberg.IcebergErrorCode.ICEBERG_INVALID_METADATA) ConnectorTableProperties(io.trino.spi.connector.ConnectorTableProperties) DiscretePredicates(io.trino.spi.connector.DiscretePredicates) ConnectorExpression(io.trino.spi.expression.ConnectorExpression) PartitionFields.parsePartitionFields(io.trino.plugin.iceberg.PartitionFields.parsePartitionFields) ArrayDeque(java.util.ArrayDeque) MaterializedViewNotFoundException(io.trino.spi.connector.MaterializedViewNotFoundException) TypeConverter.toTrinoType(io.trino.plugin.iceberg.TypeConverter.toTrinoType) IcebergUtil.getColumns(io.trino.plugin.iceberg.IcebergUtil.getColumns) IcebergSessionProperties.isProjectionPushdownEnabled(io.trino.plugin.iceberg.IcebergSessionProperties.isProjectionPushdownEnabled) IcebergTableProcedureId(io.trino.plugin.iceberg.procedure.IcebergTableProcedureId) IcebergSessionProperties.isStatisticsEnabled(io.trino.plugin.iceberg.IcebergSessionProperties.isStatisticsEnabled) AppendFiles(org.apache.iceberg.AppendFiles) NO_RETRIES(io.trino.spi.connector.RetryMode.NO_RETRIES) ConnectorMaterializedViewDefinition(io.trino.spi.connector.ConnectorMaterializedViewDefinition) TypeConverter.toIcebergType(io.trino.plugin.iceberg.TypeConverter.toIcebergType) PartitionField(org.apache.iceberg.PartitionField) DataFiles(org.apache.iceberg.DataFiles) CatalogSchemaName(io.trino.spi.connector.CatalogSchemaName) LOCATION_PROPERTY(io.trino.plugin.iceberg.IcebergTableProperties.LOCATION_PROPERTY) Path(org.apache.hadoop.fs.Path) DATA(io.trino.plugin.iceberg.TableType.DATA) ConnectorViewDefinition(io.trino.spi.connector.ConnectorViewDefinition) FileScanTask(org.apache.iceberg.FileScanTask) DataFile(org.apache.iceberg.DataFile) Splitter(com.google.common.base.Splitter) IcebergTableProperties.getPartitioning(io.trino.plugin.iceberg.IcebergTableProperties.getPartitioning) IcebergUtil.getFileFormat(io.trino.plugin.iceberg.IcebergUtil.getFileFormat) ImmutableSet(com.google.common.collect.ImmutableSet) ImmutableMap(com.google.common.collect.ImmutableMap) Collection(java.util.Collection) HiveWrittenPartitions(io.trino.plugin.hive.HiveWrittenPartitions) LocatedFileStatus(org.apache.hadoop.fs.LocatedFileStatus) ConcurrentHashMap(java.util.concurrent.ConcurrentHashMap) ComputedStatistics(io.trino.spi.statistics.ComputedStatistics) TrinoException(io.trino.spi.TrinoException) TableScan(org.apache.iceberg.TableScan) ConnectorOutputMetadata(io.trino.spi.connector.ConnectorOutputMetadata) String.format(java.lang.String.format) SchemaParser(org.apache.iceberg.SchemaParser) DataSize(io.airlift.units.DataSize) HdfsContext(io.trino.plugin.hive.HdfsEnvironment.HdfsContext) List(java.util.List) BIGINT(io.trino.spi.type.BigintType.BIGINT) ClassLoaderSafeSystemTable(io.trino.plugin.base.classloader.ClassLoaderSafeSystemTable) IcebergUtil.getTableComment(io.trino.plugin.iceberg.IcebergUtil.getTableComment) Assignment(io.trino.spi.connector.Assignment) HiveApplyProjectionUtil(io.trino.plugin.hive.HiveApplyProjectionUtil) BeginTableExecuteResult(io.trino.spi.connector.BeginTableExecuteResult) PartitionSpec(org.apache.iceberg.PartitionSpec) Function.identity(java.util.function.Function.identity) TableProperties(org.apache.iceberg.TableProperties) Optional(java.util.Optional) ProjectedColumnRepresentation(io.trino.plugin.hive.HiveApplyProjectionUtil.ProjectedColumnRepresentation) ConnectorMetadata(io.trino.spi.connector.ConnectorMetadata) Pattern(java.util.regex.Pattern) SystemTable(io.trino.spi.connector.SystemTable) JsonCodec(io.airlift.json.JsonCodec) Constraint(io.trino.spi.connector.Constraint) IcebergOptimizeHandle(io.trino.plugin.iceberg.procedure.IcebergOptimizeHandle) Logger(io.airlift.log.Logger) PartitionFields.toPartitionFields(io.trino.plugin.iceberg.PartitionFields.toPartitionFields) HashMap(java.util.HashMap) Deque(java.util.Deque) Function(java.util.function.Function) ExpressionConverter.toIcebergExpression(io.trino.plugin.iceberg.ExpressionConverter.toIcebergExpression) PARTITIONING_PROPERTY(io.trino.plugin.iceberg.IcebergTableProperties.PARTITIONING_PROPERTY) HashSet(java.util.HashSet) BiPredicate(java.util.function.BiPredicate) ImmutableList(com.google.common.collect.ImmutableList) Verify.verify(com.google.common.base.Verify.verify) Objects.requireNonNull(java.util.Objects.requireNonNull) Suppliers(com.google.common.base.Suppliers) TableStatistics(io.trino.spi.statistics.TableStatistics) HiveApplyProjectionUtil.extractSupportedProjectedColumns(io.trino.plugin.hive.HiveApplyProjectionUtil.extractSupportedProjectedColumns) VerifyException(com.google.common.base.VerifyException) RetryMode(io.trino.spi.connector.RetryMode) Iterator(java.util.Iterator) TupleDomain(io.trino.spi.predicate.TupleDomain) IcebergUtil.toIcebergSchema(io.trino.plugin.iceberg.IcebergUtil.toIcebergSchema) Transaction(org.apache.iceberg.Transaction) Comparator(java.util.Comparator) TypeManager(io.trino.spi.type.TypeManager) HiveUtil.isStructuralType(io.trino.plugin.hive.util.HiveUtil.isStructuralType) Snapshot(org.apache.iceberg.Snapshot) Variable(io.trino.spi.expression.Variable) ConcurrentHashMap(java.util.concurrent.ConcurrentHashMap) HashMap(java.util.HashMap) ConnectorExpression(io.trino.spi.expression.ConnectorExpression) ImmutableMap.toImmutableMap(com.google.common.collect.ImmutableMap.toImmutableMap) ImmutableMap(com.google.common.collect.ImmutableMap) Assignment(io.trino.spi.connector.Assignment) ImmutableSet.toImmutableSet(com.google.common.collect.ImmutableSet.toImmutableSet) ImmutableSet(com.google.common.collect.ImmutableSet) ProjectedColumnRepresentation(io.trino.plugin.hive.HiveApplyProjectionUtil.ProjectedColumnRepresentation) Map(java.util.Map) ImmutableMap.toImmutableMap(com.google.common.collect.ImmutableMap.toImmutableMap) ImmutableMap(com.google.common.collect.ImmutableMap) ConcurrentHashMap(java.util.concurrent.ConcurrentHashMap) HashMap(java.util.HashMap)

Example 5 with ProjectionApplicationResult

use of io.trino.spi.connector.ProjectionApplicationResult in project trino by trinodb.

the class PruneTableScanColumns method pruneColumns.

public static Optional<PlanNode> pruneColumns(Metadata metadata, TypeProvider types, Session session, TableScanNode node, Set<Symbol> referencedOutputs) {
    List<Symbol> newOutputs = filteredCopy(node.getOutputSymbols(), referencedOutputs::contains);
    if (newOutputs.size() == node.getOutputSymbols().size()) {
        return Optional.empty();
    }
    List<ConnectorExpression> projections = newOutputs.stream().map(symbol -> new Variable(symbol.getName(), types.get(symbol))).collect(toImmutableList());
    TableHandle handle = node.getTable();
    Optional<ProjectionApplicationResult<TableHandle>> result = metadata.applyProjection(session, handle, projections, newOutputs.stream().collect(toImmutableMap(Symbol::getName, node.getAssignments()::get)));
    Map<Symbol, ColumnHandle> newAssignments;
    // Bail out if the connector does anything other than limit the list of columns (e.g., if it synthesizes arbitrary expressions)
    if (result.isPresent() && result.get().getProjections().stream().allMatch(Variable.class::isInstance)) {
        handle = result.get().getHandle();
        Map<String, ColumnHandle> assignments = result.get().getAssignments().stream().collect(toImmutableMap(Assignment::getVariable, Assignment::getColumn));
        ImmutableMap.Builder<Symbol, ColumnHandle> builder = ImmutableMap.builder();
        for (int i = 0; i < newOutputs.size(); i++) {
            Variable variable = (Variable) result.get().getProjections().get(i);
            builder.put(newOutputs.get(i), assignments.get(variable.getName()));
        }
        newAssignments = builder.buildOrThrow();
    } else {
        newAssignments = newOutputs.stream().collect(toImmutableMap(Function.identity(), node.getAssignments()::get));
    }
    Set<ColumnHandle> visibleColumns = ImmutableSet.copyOf(newAssignments.values());
    TupleDomain<ColumnHandle> enforcedConstraint = node.getEnforcedConstraint().filter((columnHandle, domain) -> visibleColumns.contains(columnHandle));
    Optional<PlanNodeStatsEstimate> newStatistics = node.getStatistics().map(statistics -> new PlanNodeStatsEstimate(statistics.getOutputRowCount(), statistics.getSymbolStatistics().entrySet().stream().filter(entry -> newAssignments.containsKey(entry.getKey())).collect(toImmutableMap(Entry::getKey, Entry::getValue))));
    return Optional.of(new TableScanNode(node.getId(), handle, newOutputs, newAssignments, enforcedConstraint, newStatistics, node.isUpdateTarget(), node.getUseConnectorNodePartitioning()));
}
Also used : Variable(io.trino.spi.expression.Variable) Function(java.util.function.Function) PlanNode(io.trino.sql.planner.plan.PlanNode) Map(java.util.Map) Objects.requireNonNull(java.util.Objects.requireNonNull) ColumnHandle(io.trino.spi.connector.ColumnHandle) ProjectionApplicationResult(io.trino.spi.connector.ProjectionApplicationResult) TableScanNode(io.trino.sql.planner.plan.TableScanNode) PlanNodeStatsEstimate(io.trino.cost.PlanNodeStatsEstimate) Symbol(io.trino.sql.planner.Symbol) MoreLists.filteredCopy(io.trino.util.MoreLists.filteredCopy) ImmutableSet(com.google.common.collect.ImmutableSet) ImmutableMap(com.google.common.collect.ImmutableMap) ImmutableList.toImmutableList(com.google.common.collect.ImmutableList.toImmutableList) Set(java.util.Set) TupleDomain(io.trino.spi.predicate.TupleDomain) Patterns.tableScan(io.trino.sql.planner.plan.Patterns.tableScan) List(java.util.List) ImmutableMap.toImmutableMap(com.google.common.collect.ImmutableMap.toImmutableMap) TableHandle(io.trino.metadata.TableHandle) ConnectorExpression(io.trino.spi.expression.ConnectorExpression) Assignment(io.trino.spi.connector.Assignment) Entry(java.util.Map.Entry) Metadata(io.trino.metadata.Metadata) TypeProvider(io.trino.sql.planner.TypeProvider) Optional(java.util.Optional) Session(io.trino.Session) ColumnHandle(io.trino.spi.connector.ColumnHandle) Variable(io.trino.spi.expression.Variable) PlanNodeStatsEstimate(io.trino.cost.PlanNodeStatsEstimate) Symbol(io.trino.sql.planner.Symbol) ConnectorExpression(io.trino.spi.expression.ConnectorExpression) ProjectionApplicationResult(io.trino.spi.connector.ProjectionApplicationResult) ImmutableMap(com.google.common.collect.ImmutableMap) ImmutableMap.toImmutableMap(com.google.common.collect.ImmutableMap.toImmutableMap) TableScanNode(io.trino.sql.planner.plan.TableScanNode) TableHandle(io.trino.metadata.TableHandle)

Aggregations

ProjectionApplicationResult (io.trino.spi.connector.ProjectionApplicationResult)7 ImmutableList.toImmutableList (com.google.common.collect.ImmutableList.toImmutableList)6 ImmutableMap (com.google.common.collect.ImmutableMap)6 Assignment (io.trino.spi.connector.Assignment)6 ColumnHandle (io.trino.spi.connector.ColumnHandle)6 ConnectorExpression (io.trino.spi.expression.ConnectorExpression)6 Variable (io.trino.spi.expression.Variable)6 TupleDomain (io.trino.spi.predicate.TupleDomain)6 List (java.util.List)6 Map (java.util.Map)6 Optional (java.util.Optional)6 ImmutableList (com.google.common.collect.ImmutableList)5 ImmutableMap.toImmutableMap (com.google.common.collect.ImmutableMap.toImmutableMap)5 Verify.verify (com.google.common.base.Verify.verify)4 ImmutableSet (com.google.common.collect.ImmutableSet)4 ConnectorSession (io.trino.spi.connector.ConnectorSession)4 ConnectorTableHandle (io.trino.spi.connector.ConnectorTableHandle)4 ImmutableSet.toImmutableSet (com.google.common.collect.ImmutableSet.toImmutableSet)3 ColumnMetadata (io.trino.spi.connector.ColumnMetadata)3 ConnectorInsertTableHandle (io.trino.spi.connector.ConnectorInsertTableHandle)3