Examples with StorageFormat com.facebook.presto.hive.metastore.StorageFormat used on opensource projects

Example 1 with StorageFormat

use of com.facebook.presto.hive.metastore.StorageFormat in project presto by prestodb.

the class AbstractTestHiveClient method doTestMismatchSchemaTable.

protected void doTestMismatchSchemaTable(SchemaTableName schemaTableName, HiveStorageFormat storageFormat, List<ColumnMetadata> tableBefore, MaterializedResult dataBefore, List<ColumnMetadata> tableAfter, MaterializedResult dataAfter) throws Exception {
    String schemaName = schemaTableName.getSchemaName();
    String tableName = schemaTableName.getTableName();
    doCreateEmptyTable(schemaTableName, storageFormat, tableBefore);
    // insert the data
    try (Transaction transaction = newTransaction()) {
        ConnectorSession session = newSession();
        ConnectorMetadata metadata = transaction.getMetadata();
        ConnectorTableHandle tableHandle = getTableHandle(metadata, schemaTableName);
        ConnectorInsertTableHandle insertTableHandle = metadata.beginInsert(session, tableHandle);
        ConnectorPageSink sink = pageSinkProvider.createPageSink(transaction.getTransactionHandle(), session, insertTableHandle);
        sink.appendPage(dataBefore.toPage());
        Collection<Slice> fragments = getFutureValue(sink.finish());
        metadata.finishInsert(session, insertTableHandle, fragments);
        transaction.commit();
    }
    // load the table and verify the data
    try (Transaction transaction = newTransaction()) {
        ConnectorSession session = newSession();
        ConnectorMetadata metadata = transaction.getMetadata();
        ConnectorTableHandle tableHandle = getTableHandle(metadata, schemaTableName);
        List<ColumnHandle> columnHandles = metadata.getColumnHandles(session, tableHandle).values().stream().filter(columnHandle -> !((HiveColumnHandle) columnHandle).isHidden()).collect(toList());
        MaterializedResult result = readTable(transaction, tableHandle, columnHandles, session, TupleDomain.all(), OptionalInt.empty(), Optional.empty());
        assertEqualsIgnoreOrder(result.getMaterializedRows(), dataBefore.getMaterializedRows());
        transaction.commit();
    }
    // alter the table schema
    try (Transaction transaction = newTransaction()) {
        ConnectorSession session = newSession();
        PrincipalPrivileges principalPrivileges = new PrincipalPrivileges(ImmutableMultimap.<String, HivePrivilegeInfo>builder().put(session.getUser(), new HivePrivilegeInfo(HivePrivilege.SELECT, true)).put(session.getUser(), new HivePrivilegeInfo(HivePrivilege.INSERT, true)).put(session.getUser(), new HivePrivilegeInfo(HivePrivilege.UPDATE, true)).put(session.getUser(), new HivePrivilegeInfo(HivePrivilege.DELETE, true)).build(), ImmutableMultimap.of());
        Table oldTable = transaction.getMetastore(schemaName).getTable(schemaName, tableName).get();
        HiveTypeTranslator hiveTypeTranslator = new HiveTypeTranslator();
        List<Column> dataColumns = tableAfter.stream().filter(columnMetadata -> !columnMetadata.getName().equals("ds")).map(columnMetadata -> new Column(columnMetadata.getName(), toHiveType(hiveTypeTranslator, columnMetadata.getType()), Optional.empty())).collect(toList());
        Table.Builder newTable = Table.builder(oldTable).setDataColumns(dataColumns);
        transaction.getMetastore(schemaName).replaceView(schemaName, tableName, newTable.build(), principalPrivileges);
        transaction.commit();
    }
    // load the altered table and verify the data
    try (Transaction transaction = newTransaction()) {
        ConnectorSession session = newSession();
        ConnectorMetadata metadata = transaction.getMetadata();
        ConnectorTableHandle tableHandle = getTableHandle(metadata, schemaTableName);
        List<ColumnHandle> columnHandles = metadata.getColumnHandles(session, tableHandle).values().stream().filter(columnHandle -> !((HiveColumnHandle) columnHandle).isHidden()).collect(toList());
        MaterializedResult result = readTable(transaction, tableHandle, columnHandles, session, TupleDomain.all(), OptionalInt.empty(), Optional.empty());
        assertEqualsIgnoreOrder(result.getMaterializedRows(), dataAfter.getMaterializedRows());
        transaction.commit();
    }
    // insertions to the partitions with type mismatches should fail
    try (Transaction transaction = newTransaction()) {
        ConnectorSession session = newSession();
        ConnectorMetadata metadata = transaction.getMetadata();
        ConnectorTableHandle tableHandle = getTableHandle(metadata, schemaTableName);
        ConnectorInsertTableHandle insertTableHandle = metadata.beginInsert(session, tableHandle);
        ConnectorPageSink sink = pageSinkProvider.createPageSink(transaction.getTransactionHandle(), session, insertTableHandle);
        sink.appendPage(dataAfter.toPage());
        Collection<Slice> fragments = getFutureValue(sink.finish());
        metadata.finishInsert(session, insertTableHandle, fragments);
        transaction.commit();
        fail("expected exception");
    } catch (PrestoException e) {
        // expected
        assertEquals(e.getErrorCode(), HIVE_PARTITION_SCHEMA_MISMATCH.toErrorCode());
    }
}

Example 2 with StorageFormat

use of com.facebook.presto.hive.metastore.StorageFormat in project presto by prestodb.

the class RcFileFileWriterFactory method createFileWriter.

@Override
public Optional<HiveFileWriter> createFileWriter(Path path, List<String> inputColumnNames, StorageFormat storageFormat, Properties schema, JobConf configuration, ConnectorSession session, Optional<EncryptionInformation> encryptionInformation) {
    if (!HiveSessionProperties.isRcfileOptimizedWriterEnabled(session)) {
        return Optional.empty();
    }
    if (!RCFileOutputFormat.class.getName().equals(storageFormat.getOutputFormat())) {
        return Optional.empty();
    }
    RcFileEncoding rcFileEncoding;
    if (LazyBinaryColumnarSerDe.class.getName().equals(storageFormat.getSerDe())) {
        rcFileEncoding = new BinaryRcFileEncoding();
    } else if (ColumnarSerDe.class.getName().equals(storageFormat.getSerDe())) {
        rcFileEncoding = createTextVectorEncoding(schema, hiveStorageTimeZone);
    } else {
        return Optional.empty();
    }
    Optional<String> codecName = Optional.ofNullable(configuration.get(FileOutputFormat.COMPRESS_CODEC));
    // existing tables and partitions may have columns in a different order than the writer is providing, so build
    // an index to rearrange columns in the proper order
    List<String> fileColumnNames = Splitter.on(',').trimResults().omitEmptyStrings().splitToList(schema.getProperty(META_TABLE_COLUMNS, ""));
    List<Type> fileColumnTypes = toHiveTypes(schema.getProperty(META_TABLE_COLUMN_TYPES, "")).stream().map(hiveType -> hiveType.getType(typeManager)).collect(toList());
    int[] fileInputColumnIndexes = fileColumnNames.stream().mapToInt(inputColumnNames::indexOf).toArray();
    try {
        FileSystem fileSystem = hdfsEnvironment.getFileSystem(session.getUser(), path, configuration);
        OutputStream outputStream = fileSystem.create(path);
        Optional<Supplier<RcFileDataSource>> validationInputFactory = Optional.empty();
        if (HiveSessionProperties.isRcfileOptimizedWriterValidate(session)) {
            validationInputFactory = Optional.of(() -> {
                try {
                    return new HdfsRcFileDataSource(path.toString(), fileSystem.open(path), fileSystem.getFileStatus(path).getLen(), stats);
                } catch (IOException e) {
                    throw new PrestoException(HIVE_WRITE_VALIDATION_FAILED, e);
                }
            });
        }
        Callable<Void> rollbackAction = () -> {
            fileSystem.delete(path, false);
            return null;
        };
        return Optional.of(new RcFileFileWriter(outputStream, rollbackAction, rcFileEncoding, fileColumnTypes, codecName, fileInputColumnIndexes, ImmutableMap.<String, String>builder().put(HiveMetadata.PRESTO_VERSION_NAME, nodeVersion.toString()).put(MetastoreUtil.PRESTO_QUERY_ID_NAME, session.getQueryId()).build(), validationInputFactory));
    } catch (Exception e) {
        throw new PrestoException(HIVE_WRITER_OPEN_ERROR, "Error creating RCFile file", e);
    }
}

Example 3 with StorageFormat

use of com.facebook.presto.hive.metastore.StorageFormat in project presto by prestodb.

the class AbstractTestHiveClient method doTestTransactionDeleteInsert.

private void doTestTransactionDeleteInsert(HiveStorageFormat storageFormat, SchemaTableName tableName, Domain domainToDrop, MaterializedResult insertData, MaterializedResult expectedData, TransactionDeleteInsertTestTag tag, boolean expectQuerySucceed, Optional<ConflictTrigger> conflictTrigger) throws Exception {
    Path writePath = null;
    Path targetPath = null;
    try (Transaction transaction = newTransaction()) {
        try {
            ConnectorMetadata metadata = transaction.getMetadata();
            ConnectorTableHandle tableHandle = getTableHandle(metadata, tableName);
            ConnectorSession session;
            rollbackIfEquals(tag, ROLLBACK_RIGHT_AWAY);
            // Query 1: delete
            session = newSession();
            HiveColumnHandle dsColumnHandle = (HiveColumnHandle) metadata.getColumnHandles(session, tableHandle).get("pk2");
            TupleDomain<ColumnHandle> tupleDomain = withColumnDomains(ImmutableMap.of(dsColumnHandle, domainToDrop));
            Constraint<ColumnHandle> constraint = new Constraint<>(tupleDomain, convertToPredicate(tupleDomain));
            ConnectorTableLayoutHandle tableLayoutHandle = getTableLayout(session, metadata, tableHandle, constraint, transaction).getHandle();
            metadata.metadataDelete(session, tableHandle, tableLayoutHandle);
            rollbackIfEquals(tag, ROLLBACK_AFTER_DELETE);
            // Query 2: insert
            session = newSession();
            ConnectorInsertTableHandle insertTableHandle = metadata.beginInsert(session, tableHandle);
            rollbackIfEquals(tag, ROLLBACK_AFTER_BEGIN_INSERT);
            writePath = getStagingPathRoot(insertTableHandle);
            targetPath = getTargetPathRoot(insertTableHandle);
            ConnectorPageSink sink = pageSinkProvider.createPageSink(transaction.getTransactionHandle(), session, insertTableHandle, TEST_HIVE_PAGE_SINK_CONTEXT);
            sink.appendPage(insertData.toPage());
            rollbackIfEquals(tag, ROLLBACK_AFTER_APPEND_PAGE);
            Collection<Slice> fragments = getFutureValue(sink.finish());
            rollbackIfEquals(tag, ROLLBACK_AFTER_SINK_FINISH);
            metadata.finishInsert(session, insertTableHandle, fragments, ImmutableList.of());
            rollbackIfEquals(tag, ROLLBACK_AFTER_FINISH_INSERT);
            assertEquals(tag, COMMIT);
            if (conflictTrigger.isPresent()) {
                List<PartitionUpdate> partitionUpdates = fragments.stream().map(Slice::getBytes).map(HiveTestUtils.PARTITION_UPDATE_CODEC::fromJson).collect(toList());
                conflictTrigger.get().triggerConflict(session, tableName, insertTableHandle, partitionUpdates);
            }
            transaction.commit();
            if (conflictTrigger.isPresent()) {
                assertTrue(expectQuerySucceed);
                conflictTrigger.get().verifyAndCleanup(tableName);
            }
        } catch (TestingRollbackException e) {
            transaction.rollback();
        } catch (PrestoException e) {
            assertFalse(expectQuerySucceed);
            if (conflictTrigger.isPresent()) {
                conflictTrigger.get().verifyAndCleanup(tableName);
            }
        }
    }
    // check that temporary files are removed
    if (writePath != null && !writePath.equals(targetPath)) {
        HdfsContext context = new HdfsContext(newSession(), tableName.getSchemaName(), tableName.getTableName(), writePath.toString(), false);
        FileSystem fileSystem = hdfsEnvironment.getFileSystem(context, writePath);
        assertFalse(fileSystem.exists(writePath));
    }
    try (Transaction transaction = newTransaction()) {
        ConnectorSession session = newSession();
        MetastoreContext metastoreContext = new MetastoreContext(session.getIdentity(), session.getQueryId(), session.getClientInfo(), session.getSource(), getMetastoreHeaders(session), false, DEFAULT_COLUMN_CONVERTER_PROVIDER);
        // verify partitions
        List<String> partitionNames = transaction.getMetastore().getPartitionNames(metastoreContext, tableName.getSchemaName(), tableName.getTableName()).orElseThrow(() -> new AssertionError("Table does not exist: " + tableName));
        assertEqualsIgnoreOrder(partitionNames, expectedData.getMaterializedRows().stream().map(row -> format("pk1=%s/pk2=%s", row.getField(1), row.getField(2))).distinct().collect(toList()));
        // load the new table
        ConnectorMetadata metadata = transaction.getMetadata();
        ConnectorTableHandle tableHandle = getTableHandle(metadata, tableName);
        List<ColumnHandle> columnHandles = filterNonHiddenColumnHandles(metadata.getColumnHandles(session, tableHandle).values());
        // verify the data
        MaterializedResult result = readTable(transaction, tableHandle, columnHandles, session, TupleDomain.all(), OptionalInt.empty(), Optional.of(storageFormat));
        assertEqualsIgnoreOrder(result.getMaterializedRows(), expectedData.getMaterializedRows());
    }
}

Example 4 with StorageFormat

use of com.facebook.presto.hive.metastore.StorageFormat in project presto by prestodb.

the class AbstractTestHiveClient method doCreateEmptyTable.

protected void doCreateEmptyTable(SchemaTableName tableName, HiveStorageFormat storageFormat, List<ColumnMetadata> createTableColumns) throws Exception {
    List<String> partitionedBy = createTableColumns.stream().filter(column -> column.getName().equals("ds")).map(ColumnMetadata::getName).collect(toList());
    String queryId;
    try (Transaction transaction = newTransaction()) {
        ConnectorSession session = newSession();
        ConnectorMetadata metadata = transaction.getMetadata();
        queryId = session.getQueryId();
        ConnectorTableMetadata tableMetadata = new ConnectorTableMetadata(tableName, createTableColumns, createTableProperties(storageFormat, partitionedBy));
        metadata.createTable(session, tableMetadata, false);
        transaction.commit();
    }
    try (Transaction transaction = newTransaction()) {
        ConnectorSession session = newSession();
        ConnectorMetadata metadata = transaction.getMetadata();
        MetastoreContext metastoreContext = new MetastoreContext(session.getIdentity(), session.getQueryId(), session.getClientInfo(), session.getSource(), getMetastoreHeaders(session), false, DEFAULT_COLUMN_CONVERTER_PROVIDER);
        // load the new table
        ConnectorTableHandle tableHandle = getTableHandle(metadata, tableName);
        // verify the metadata
        ConnectorTableMetadata tableMetadata = metadata.getTableMetadata(session, getTableHandle(metadata, tableName));
        List<ColumnMetadata> expectedColumns = createTableColumns.stream().map(column -> new ColumnMetadata(column.getName(), column.getType(), column.getComment(), columnExtraInfo(partitionedBy.contains(column.getName())), false)).collect(toList());
        assertEquals(filterNonHiddenColumnMetadata(tableMetadata.getColumns()), expectedColumns);
        // verify table format
        Table table = transaction.getMetastore().getTable(metastoreContext, tableName.getSchemaName(), tableName.getTableName()).get();
        assertEquals(table.getStorage().getStorageFormat().getInputFormat(), storageFormat.getInputFormat());
        // verify the node version and query ID
        assertEquals(table.getParameters().get(PRESTO_VERSION_NAME), TEST_SERVER_VERSION);
        assertEquals(table.getParameters().get(PRESTO_QUERY_ID_NAME), queryId);
        // verify the table is empty
        List<ColumnHandle> columnHandles = filterNonHiddenColumnHandles(metadata.getColumnHandles(session, tableHandle).values());
        MaterializedResult result = readTable(transaction, tableHandle, columnHandles, session, TupleDomain.all(), OptionalInt.empty(), Optional.of(storageFormat));
        assertEquals(result.getRowCount(), 0);
        // verify basic statistics
        if (partitionedBy.isEmpty()) {
            assertEmptyTableStatistics(metastoreContext, tableName, transaction);
        }
    }
}

Example 5 with StorageFormat

use of com.facebook.presto.hive.metastore.StorageFormat in project presto by prestodb.

the class AbstractTestHiveClient method doTestMetadataDelete.

private void doTestMetadataDelete(HiveStorageFormat storageFormat, SchemaTableName tableName) throws Exception {
    // creating the table
    doCreateEmptyTable(tableName, storageFormat, CREATE_TABLE_COLUMNS_PARTITIONED);
    insertData(tableName, CREATE_TABLE_PARTITIONED_DATA);
    MaterializedResult.Builder expectedResultBuilder = MaterializedResult.resultBuilder(SESSION, CREATE_TABLE_PARTITIONED_DATA.getTypes());
    expectedResultBuilder.rows(CREATE_TABLE_PARTITIONED_DATA.getMaterializedRows());
    try (Transaction transaction = newTransaction()) {
        ConnectorSession session = newSession();
        ConnectorMetadata metadata = transaction.getMetadata();
        MetastoreContext metastoreContext = new MetastoreContext(session.getIdentity(), session.getQueryId(), session.getClientInfo(), session.getSource(), getMetastoreHeaders(session), false, DEFAULT_COLUMN_CONVERTER_PROVIDER);
        // verify partitions were created
        List<String> partitionNames = transaction.getMetastore().getPartitionNames(metastoreContext, tableName.getSchemaName(), tableName.getTableName()).orElseThrow(() -> new AssertionError("Table does not exist: " + tableName));
        assertEqualsIgnoreOrder(partitionNames, CREATE_TABLE_PARTITIONED_DATA.getMaterializedRows().stream().map(row -> "ds=" + row.getField(CREATE_TABLE_PARTITIONED_DATA.getTypes().size() - 1)).collect(toList()));
        // verify table directory is not empty
        Set<String> filesAfterInsert = listAllDataFiles(metastoreContext, transaction, tableName.getSchemaName(), tableName.getTableName());
        assertFalse(filesAfterInsert.isEmpty());
        // verify the data
        ConnectorTableHandle tableHandle = getTableHandle(metadata, tableName);
        List<ColumnHandle> columnHandles = filterNonHiddenColumnHandles(metadata.getColumnHandles(session, tableHandle).values());
        MaterializedResult result = readTable(transaction, tableHandle, columnHandles, session, TupleDomain.all(), OptionalInt.empty(), Optional.of(storageFormat));
        assertEqualsIgnoreOrder(result.getMaterializedRows(), expectedResultBuilder.build().getMaterializedRows());
    }
    try (Transaction transaction = newTransaction()) {
        ConnectorSession session = newSession();
        ConnectorMetadata metadata = transaction.getMetadata();
        // get ds column handle
        ConnectorTableHandle tableHandle = getTableHandle(metadata, tableName);
        HiveColumnHandle dsColumnHandle = (HiveColumnHandle) metadata.getColumnHandles(session, tableHandle).get("ds");
        // delete ds=2015-07-03
        session = newSession();
        TupleDomain<ColumnHandle> tupleDomain = TupleDomain.fromFixedValues(ImmutableMap.of(dsColumnHandle, NullableValue.of(createUnboundedVarcharType(), utf8Slice("2015-07-03"))));
        Constraint<ColumnHandle> constraint = new Constraint<>(tupleDomain, convertToPredicate(tupleDomain));
        ConnectorTableLayoutHandle tableLayoutHandle = getTableLayout(session, metadata, tableHandle, constraint, transaction).getHandle();
        metadata.metadataDelete(session, tableHandle, tableLayoutHandle);
        transaction.commit();
    }
    try (Transaction transaction = newTransaction()) {
        ConnectorSession session = newSession();
        ConnectorMetadata metadata = transaction.getMetadata();
        ConnectorTableHandle tableHandle = getTableHandle(metadata, tableName);
        List<ColumnHandle> columnHandles = filterNonHiddenColumnHandles(metadata.getColumnHandles(session, tableHandle).values());
        HiveColumnHandle dsColumnHandle = (HiveColumnHandle) metadata.getColumnHandles(session, tableHandle).get("ds");
        int dsColumnOrdinalPosition = columnHandles.indexOf(dsColumnHandle);
        // verify the data
        session = newSession();
        ImmutableList<MaterializedRow> expectedRows = expectedResultBuilder.build().getMaterializedRows().stream().filter(row -> !"2015-07-03".equals(row.getField(dsColumnOrdinalPosition))).collect(toImmutableList());
        MaterializedResult actualAfterDelete = readTable(transaction, tableHandle, columnHandles, session, TupleDomain.all(), OptionalInt.empty(), Optional.of(storageFormat));
        assertEqualsIgnoreOrder(actualAfterDelete.getMaterializedRows(), expectedRows);
    }
    try (Transaction transaction = newTransaction()) {
        ConnectorSession session = newSession();
        ConnectorMetadata metadata = transaction.getMetadata();
        ConnectorTableHandle tableHandle = getTableHandle(metadata, tableName);
        HiveColumnHandle dsColumnHandle = (HiveColumnHandle) metadata.getColumnHandles(session, tableHandle).get("ds");
        // delete ds=2015-07-01 and 2015-07-02
        session = newSession();
        TupleDomain<ColumnHandle> tupleDomain2 = withColumnDomains(ImmutableMap.of(dsColumnHandle, Domain.create(ValueSet.ofRanges(Range.range(createUnboundedVarcharType(), utf8Slice("2015-07-01"), true, utf8Slice("2015-07-02"), true)), false)));
        Constraint<ColumnHandle> constraint2 = new Constraint<>(tupleDomain2, convertToPredicate(tupleDomain2));
        ConnectorTableLayoutHandle tableLayoutHandle2 = getTableLayout(session, metadata, tableHandle, constraint2, transaction).getHandle();
        metadata.metadataDelete(session, tableHandle, tableLayoutHandle2);
        transaction.commit();
    }
    try (Transaction transaction = newTransaction()) {
        ConnectorSession session = newSession();
        ConnectorMetadata metadata = transaction.getMetadata();
        ConnectorTableHandle tableHandle = getTableHandle(metadata, tableName);
        List<ColumnHandle> columnHandles = ImmutableList.copyOf(metadata.getColumnHandles(session, tableHandle).values());
        // verify the data
        session = newSession();
        MaterializedResult actualAfterDelete2 = readTable(transaction, tableHandle, columnHandles, session, TupleDomain.all(), OptionalInt.empty(), Optional.of(storageFormat));
        assertEqualsIgnoreOrder(actualAfterDelete2.getMaterializedRows(), ImmutableList.of());
        // verify table directory is empty
        MetastoreContext metastoreContext = new MetastoreContext(session.getIdentity(), session.getQueryId(), session.getClientInfo(), session.getSource(), getMetastoreHeaders(session), false, DEFAULT_COLUMN_CONVERTER_PROVIDER);
        Set<String> filesAfterDelete = listAllDataFiles(metastoreContext, transaction, tableName.getSchemaName(), tableName.getTableName());
        assertTrue(filesAfterDelete.isEmpty());
    }
}