@Override
public void createTable(ConnectorSession session, ConnectorTableMetadata tableMetadata) {
checkArgument(!isNullOrEmpty(tableMetadata.getOwner()), "Table owner is null or empty");
SchemaTableName schemaTableName = tableMetadata.getTable();
String schemaName = schemaTableName.getSchemaName();
String tableName = schemaTableName.getTableName();
ImmutableList.Builder<String> columnNames = ImmutableList.builder();
ImmutableList.Builder<Type> columnTypes = ImmutableList.builder();
buildColumnInfo(tableMetadata, columnNames, columnTypes);
ImmutableList.Builder<FieldSchema> partitionKeys = ImmutableList.builder();
ImmutableList.Builder<FieldSchema> columns = ImmutableList.builder();
List<String> names = columnNames.build();
List<String> typeNames =
columnTypes
.build()
.stream()
.map(HiveType::toHiveType)
.map(HiveType::getHiveTypeName)
.collect(toList());
for (int i = 0; i < names.size(); i++) {
if (tableMetadata.getColumns().get(i).isPartitionKey()) {
partitionKeys.add(new FieldSchema(names.get(i), typeNames.get(i), null));
} else {
columns.add(new FieldSchema(names.get(i), typeNames.get(i), null));
}
}
Path targetPath = getTargetPath(schemaName, tableName, schemaTableName);
HiveStorageFormat hiveStorageFormat = getHiveStorageFormat(session, this.hiveStorageFormat);
SerDeInfo serdeInfo = new SerDeInfo();
serdeInfo.setName(tableName);
serdeInfo.setSerializationLib(hiveStorageFormat.getSerDe());
StorageDescriptor sd = new StorageDescriptor();
sd.setLocation(targetPath.toString());
sd.setCols(columns.build());
sd.setSerdeInfo(serdeInfo);
sd.setInputFormat(hiveStorageFormat.getInputFormat());
sd.setOutputFormat(hiveStorageFormat.getOutputFormat());
Table table = new Table();
table.setDbName(schemaName);
table.setTableName(tableName);
table.setOwner(tableMetadata.getOwner());
table.setTableType(TableType.MANAGED_TABLE.toString());
String tableComment = "Created by Presto";
table.setParameters(ImmutableMap.of("comment", tableComment));
table.setPartitionKeys(partitionKeys.build());
table.setSd(sd);
metastore.createTable(table);
}
private static StorageDescriptor makeStorageDescriptor(
String tableName, List<Column> columns, Storage storage) {
if (storage.isSorted() || storage.isSkewed()) {
throw new IllegalArgumentException(
"Writing to sorted and/or skewed table/partition is not supported");
}
SerDeInfo serdeInfo = new SerDeInfo();
serdeInfo.setName(tableName);
serdeInfo.setSerializationLib(storage.getStorageFormat().getSerDeNullable());
serdeInfo.setParameters(storage.getSerdeParameters());
StorageDescriptor sd = new StorageDescriptor();
sd.setLocation(emptyToNull(storage.getLocation()));
sd.setCols(columns.stream().map(MetastoreUtil::toMetastoreApiFieldSchema).collect(toList()));
sd.setSerdeInfo(serdeInfo);
sd.setInputFormat(storage.getStorageFormat().getInputFormatNullable());
sd.setOutputFormat(storage.getStorageFormat().getOutputFormatNullable());
sd.setParameters(ImmutableMap.of());
Optional<HiveBucketProperty> bucketProperty = storage.getBucketProperty();
if (bucketProperty.isPresent()) {
sd.setNumBuckets(bucketProperty.get().getBucketCount());
sd.setBucketCols(bucketProperty.get().getBucketedBy());
}
return sd;
}
@Before
public void createTable() throws Exception {
// Use Junit's Assume to skip running this fixture against any storage formats whose
// SerDe is in the disabled serdes list.
Assume.assumeTrue(!DISABLED_SERDES.contains(serdeClass));
String databaseName = (dbName == null) ? MetaStoreUtils.DEFAULT_DATABASE_NAME : dbName;
try {
client.dropTable(databaseName, tableName);
} catch (Exception e) {
// Can fail with NoSuchObjectException.
}
Table tbl = new Table();
tbl.setDbName(databaseName);
tbl.setTableName(tableName);
if (isTableExternal()) {
tbl.setTableType(TableType.EXTERNAL_TABLE.toString());
} else {
tbl.setTableType(TableType.MANAGED_TABLE.toString());
}
StorageDescriptor sd = new StorageDescriptor();
sd.setCols(getTableColumns());
tbl.setPartitionKeys(getPartitionKeys());
tbl.setSd(sd);
sd.setBucketCols(new ArrayList<String>(2));
sd.setSerdeInfo(new SerDeInfo());
sd.getSerdeInfo().setName(tbl.getTableName());
sd.getSerdeInfo().setParameters(new HashMap<String, String>());
sd.getSerdeInfo().getParameters().put(serdeConstants.SERIALIZATION_FORMAT, "1");
if (isTableExternal()) {
sd.getSerdeInfo().getParameters().put("EXTERNAL", "TRUE");
}
sd.getSerdeInfo().setSerializationLib(serdeClass);
sd.setInputFormat(inputFormatClass);
sd.setOutputFormat(outputFormatClass);
Map<String, String> tableParams = new HashMap<String, String>();
if (isTableExternal()) {
tableParams.put("EXTERNAL", "TRUE");
}
if (isTableImmutable()) {
tableParams.put(hive_metastoreConstants.IS_IMMUTABLE, "true");
}
StatsSetupConst.setBasicStatsState(tableParams, StatsSetupConst.TRUE);
tbl.setParameters(tableParams);
client.createTable(tbl);
}
private static void createTable(String tableName, String tablePerm) throws Exception {
Table tbl = new Table();
tbl.setDbName(DATABASE);
tbl.setTableName(tableName);
StorageDescriptor sd = new StorageDescriptor();
sd.setCols(ColumnHolder.colMapping.get(tableName));
tbl.setSd(sd);
sd.setParameters(new HashMap<String, String>());
sd.setSerdeInfo(new SerDeInfo());
sd.getSerdeInfo().setName(tbl.getTableName());
sd.getSerdeInfo().setParameters(new HashMap<String, String>());
sd.setInputFormat(org.apache.hadoop.hive.ql.io.RCFileInputFormat.class.getName());
sd.setOutputFormat(org.apache.hadoop.hive.ql.io.RCFileOutputFormat.class.getName());
sd.getSerdeInfo().getParameters().put(serdeConstants.SERIALIZATION_FORMAT, "1");
sd.getSerdeInfo()
.setSerializationLib(org.apache.hadoop.hive.serde2.columnar.ColumnarSerDe.class.getName());
tbl.setPartitionKeys(ColumnHolder.partitionCols);
hmsc.createTable(tbl);
FileSystem fs = FileSystem.get(mrConf);
fs.setPermission(new Path(warehousedir, tableName), new FsPermission(tablePerm));
}
/**
* Tests that Impala is able to create an HdfsStorageDescriptor using all combinations of Parquet
* SerDe class name + input/output format class name.
*/
@Test
public void testParquetFileFormat()
throws DatabaseNotFoundException, InvalidStorageDescriptorException {
String[] parquetSerDe =
new String[] {
"org.apache.hadoop.hive.ql.io.parquet.serde.ParquetHiveSerDe",
"parquet.hive.serde.ParquetHiveSerDe"
};
String[] inputFormats =
new String[] {
"com.cloudera.impala.hive.serde.ParquetInputFormat",
"parquet.hive.DeprecatedParquetInputFormat",
"parquet.hive.MapredParquetInputFormat",
"org.apache.hadoop.hive.ql.io.parquet.MapredParquetInputFormat"
};
String[] outputFormats =
new String[] {
"com.cloudera.impala.hive.serde.ParquetOutputFormat",
"parquet.hive.DeprecatedParquetOutputFormat",
"parquet.hive.MapredParquetOutputFormat",
"org.apache.hadoop.hive.ql.io.parquet.MapredParquetOutputFormat"
};
for (String serDe : parquetSerDe) {
SerDeInfo serDeInfo = new SerDeInfo();
serDeInfo.setSerializationLib(serDe);
serDeInfo.setParameters(new HashMap<String, String>());
for (String inputFormat : inputFormats) {
for (String outputFormat : outputFormats) {
StorageDescriptor sd = new StorageDescriptor();
sd.setSerdeInfo(serDeInfo);
sd.setInputFormat(inputFormat);
sd.setOutputFormat(outputFormat);
assertNotNull(HdfsStorageDescriptor.fromStorageDescriptor("fakeTblName", sd));
}
}
}
}
@Override
public void commitCreateTable(
ConnectorOutputTableHandle tableHandle, Collection<Slice> fragments) {
HiveOutputTableHandle handle =
checkType(tableHandle, HiveOutputTableHandle.class, "tableHandle");
// verify no one raced us to create the target directory
Path targetPath = new Path(handle.getTargetPath());
// rename if using a temporary directory
if (handle.hasTemporaryPath()) {
if (pathExists(targetPath)) {
SchemaTableName table = new SchemaTableName(handle.getSchemaName(), handle.getTableName());
throw new PrestoException(
HIVE_PATH_ALREADY_EXISTS,
format(
"Unable to commit creation of table '%s': target directory already exists: %s",
table, targetPath));
}
// rename the temporary directory to the target
rename(new Path(handle.getTemporaryPath()), targetPath);
}
// create the table in the metastore
List<String> types =
handle
.getColumnTypes()
.stream()
.map(HiveType::toHiveType)
.map(HiveType::getHiveTypeName)
.collect(toList());
boolean sampled = false;
ImmutableList.Builder<FieldSchema> columns = ImmutableList.builder();
for (int i = 0; i < handle.getColumnNames().size(); i++) {
String name = handle.getColumnNames().get(i);
String type = types.get(i);
if (name.equals(SAMPLE_WEIGHT_COLUMN_NAME)) {
columns.add(new FieldSchema(name, type, "Presto sample weight column"));
sampled = true;
} else {
columns.add(new FieldSchema(name, type, null));
}
}
HiveStorageFormat hiveStorageFormat = handle.getHiveStorageFormat();
SerDeInfo serdeInfo = new SerDeInfo();
serdeInfo.setName(handle.getTableName());
serdeInfo.setSerializationLib(hiveStorageFormat.getSerDe());
serdeInfo.setParameters(ImmutableMap.<String, String>of());
StorageDescriptor sd = new StorageDescriptor();
sd.setLocation(targetPath.toString());
sd.setCols(columns.build());
sd.setSerdeInfo(serdeInfo);
sd.setInputFormat(hiveStorageFormat.getInputFormat());
sd.setOutputFormat(hiveStorageFormat.getOutputFormat());
sd.setParameters(ImmutableMap.<String, String>of());
Table table = new Table();
table.setDbName(handle.getSchemaName());
table.setTableName(handle.getTableName());
table.setOwner(handle.getTableOwner());
table.setTableType(TableType.MANAGED_TABLE.toString());
String tableComment = "Created by Presto";
if (sampled) {
tableComment =
"Sampled table created by Presto. Only query this table from Hive if you understand how Presto implements sampling.";
}
table.setParameters(ImmutableMap.of("comment", tableComment));
table.setPartitionKeys(ImmutableList.<FieldSchema>of());
table.setSd(sd);
metastore.createTable(table);
}
public void setFieldValue(_Fields field, Object value) {
switch (field) {
case COLS:
if (value == null) {
unsetCols();
} else {
setCols((List<FieldSchema>) value);
}
break;
case LOCATION:
if (value == null) {
unsetLocation();
} else {
setLocation((String) value);
}
break;
case INPUT_FORMAT:
if (value == null) {
unsetInputFormat();
} else {
setInputFormat((String) value);
}
break;
case OUTPUT_FORMAT:
if (value == null) {
unsetOutputFormat();
} else {
setOutputFormat((String) value);
}
break;
case COMPRESSED:
if (value == null) {
unsetCompressed();
} else {
setCompressed((Boolean) value);
}
break;
case NUM_BUCKETS:
if (value == null) {
unsetNumBuckets();
} else {
setNumBuckets((Integer) value);
}
break;
case SERDE_INFO:
if (value == null) {
unsetSerdeInfo();
} else {
setSerdeInfo((SerDeInfo) value);
}
break;
case BUCKET_COLS:
if (value == null) {
unsetBucketCols();
} else {
setBucketCols((List<String>) value);
}
break;
case SORT_COLS:
if (value == null) {
unsetSortCols();
} else {
setSortCols((List<Order>) value);
}
break;
case PARAMETERS:
if (value == null) {
unsetParameters();
} else {
setParameters((Map<String, String>) value);
}
break;
}
}
@Override
public final void createTable(final CatalogProtos.TableDescProto tableDescProto)
throws CatalogException {
HiveCatalogStoreClientPool.HiveCatalogStoreClient client = null;
TableDesc tableDesc = new TableDesc(tableDescProto);
String[] splitted = CatalogUtil.splitFQTableName(tableDesc.getName());
String databaseName = splitted[0];
String tableName = splitted[1];
try {
client = clientPool.getClient();
org.apache.hadoop.hive.metastore.api.Table table =
new org.apache.hadoop.hive.metastore.api.Table();
table.setDbName(databaseName);
table.setTableName(tableName);
table.setParameters(
new HashMap<String, String>(tableDesc.getMeta().getOptions().getAllKeyValus()));
// TODO: set owner
// table.setOwner();
StorageDescriptor sd = new StorageDescriptor();
sd.setSerdeInfo(new SerDeInfo());
sd.getSerdeInfo().setParameters(new HashMap<String, String>());
sd.getSerdeInfo().setName(table.getTableName());
// if tajo set location method, thrift client make exception as follows:
// Caused by: MetaException(message:java.lang.NullPointerException)
// If you want to modify table path, you have to modify on Hive cli.
if (tableDesc.isExternal()) {
table.setTableType(TableType.EXTERNAL_TABLE.name());
table.putToParameters("EXTERNAL", "TRUE");
Path tablePath = new Path(tableDesc.getUri());
FileSystem fs = tablePath.getFileSystem(conf);
if (fs.isFile(tablePath)) {
LOG.warn("A table path is a file, but HiveCatalogStore does not allow a file path.");
sd.setLocation(tablePath.getParent().toString());
} else {
sd.setLocation(tablePath.toString());
}
}
// set column information
List<Column> columns = tableDesc.getSchema().getRootColumns();
ArrayList<FieldSchema> cols = new ArrayList<FieldSchema>(columns.size());
for (Column eachField : columns) {
cols.add(
new FieldSchema(
eachField.getSimpleName(),
HiveCatalogUtil.getHiveFieldType(eachField.getDataType()),
""));
}
sd.setCols(cols);
// set partition keys
if (tableDesc.hasPartition()
&& tableDesc.getPartitionMethod().getPartitionType().equals(PartitionType.COLUMN)) {
List<FieldSchema> partitionKeys = new ArrayList<FieldSchema>();
for (Column eachPartitionKey :
tableDesc.getPartitionMethod().getExpressionSchema().getRootColumns()) {
partitionKeys.add(
new FieldSchema(
eachPartitionKey.getSimpleName(),
HiveCatalogUtil.getHiveFieldType(eachPartitionKey.getDataType()),
""));
}
table.setPartitionKeys(partitionKeys);
}
if (tableDesc.getMeta().getStoreType().equalsIgnoreCase(BuiltinStorages.RCFILE)) {
String serde = tableDesc.getMeta().getOption(StorageConstants.RCFILE_SERDE);
sd.setInputFormat(org.apache.hadoop.hive.ql.io.RCFileInputFormat.class.getName());
sd.setOutputFormat(org.apache.hadoop.hive.ql.io.RCFileOutputFormat.class.getName());
if (StorageConstants.DEFAULT_TEXT_SERDE.equals(serde)) {
sd.getSerdeInfo()
.setSerializationLib(
org.apache.hadoop.hive.serde2.columnar.ColumnarSerDe.class.getName());
} else {
sd.getSerdeInfo()
.setSerializationLib(
org.apache.hadoop.hive.serde2.columnar.LazyBinaryColumnarSerDe.class.getName());
}
if (tableDesc.getMeta().getOptions().containsKey(StorageConstants.RCFILE_NULL)) {
table.putToParameters(
serdeConstants.SERIALIZATION_NULL_FORMAT,
StringEscapeUtils.unescapeJava(
tableDesc.getMeta().getOption(StorageConstants.RCFILE_NULL)));
}
} else if (tableDesc.getMeta().getStoreType().equals(BuiltinStorages.TEXT)) {
sd.getSerdeInfo()
.setSerializationLib(
org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe.class.getName());
sd.setInputFormat(org.apache.hadoop.mapred.TextInputFormat.class.getName());
sd.setOutputFormat(
org.apache.hadoop.hive.ql.io.HiveIgnoreKeyTextOutputFormat.class.getName());
String fieldDelimiter =
tableDesc
.getMeta()
.getOption(
StorageConstants.TEXT_DELIMITER, StorageConstants.DEFAULT_FIELD_DELIMITER);
// User can use an unicode for filed delimiter such as \u0001, \001.
// In this case, java console will convert this value into "\\u001".
// And hive will un-espace this value again.
// As a result, user can use right field delimiter.
// So, we have to un-escape this value.
sd.getSerdeInfo()
.putToParameters(
serdeConstants.SERIALIZATION_FORMAT,
StringEscapeUtils.unescapeJava(fieldDelimiter));
sd.getSerdeInfo()
.putToParameters(
serdeConstants.FIELD_DELIM, StringEscapeUtils.unescapeJava(fieldDelimiter));
table.getParameters().remove(StorageConstants.TEXT_DELIMITER);
if (tableDesc.getMeta().containsOption(StorageConstants.TEXT_NULL)) {
table.putToParameters(
serdeConstants.SERIALIZATION_NULL_FORMAT,
StringEscapeUtils.unescapeJava(
tableDesc.getMeta().getOption(StorageConstants.TEXT_NULL)));
table.getParameters().remove(StorageConstants.TEXT_NULL);
}
} else if (tableDesc
.getMeta()
.getStoreType()
.equalsIgnoreCase(BuiltinStorages.SEQUENCE_FILE)) {
String serde = tableDesc.getMeta().getOption(StorageConstants.SEQUENCEFILE_SERDE);
sd.setInputFormat(org.apache.hadoop.mapred.SequenceFileInputFormat.class.getName());
sd.setOutputFormat(
org.apache.hadoop.hive.ql.io.HiveSequenceFileOutputFormat.class.getName());
if (StorageConstants.DEFAULT_TEXT_SERDE.equals(serde)) {
sd.getSerdeInfo()
.setSerializationLib(
org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe.class.getName());
String fieldDelimiter =
tableDesc
.getMeta()
.getOption(
StorageConstants.SEQUENCEFILE_DELIMITER,
StorageConstants.DEFAULT_FIELD_DELIMITER);
// User can use an unicode for filed delimiter such as \u0001, \001.
// In this case, java console will convert this value into "\\u001".
// And hive will un-espace this value again.
// As a result, user can use right field delimiter.
// So, we have to un-escape this value.
sd.getSerdeInfo()
.putToParameters(
serdeConstants.SERIALIZATION_FORMAT,
StringEscapeUtils.unescapeJava(fieldDelimiter));
sd.getSerdeInfo()
.putToParameters(
serdeConstants.FIELD_DELIM, StringEscapeUtils.unescapeJava(fieldDelimiter));
table.getParameters().remove(StorageConstants.SEQUENCEFILE_DELIMITER);
} else {
sd.getSerdeInfo()
.setSerializationLib(
org.apache.hadoop.hive.serde2.lazybinary.LazyBinarySerDe.class.getName());
}
if (tableDesc.getMeta().containsOption(StorageConstants.SEQUENCEFILE_NULL)) {
table.putToParameters(
serdeConstants.SERIALIZATION_NULL_FORMAT,
StringEscapeUtils.unescapeJava(
tableDesc.getMeta().getOption(StorageConstants.SEQUENCEFILE_NULL)));
table.getParameters().remove(StorageConstants.SEQUENCEFILE_NULL);
}
} else {
if (tableDesc.getMeta().getStoreType().equalsIgnoreCase(BuiltinStorages.PARQUET)) {
sd.setInputFormat(parquet.hive.DeprecatedParquetInputFormat.class.getName());
sd.setOutputFormat(parquet.hive.DeprecatedParquetOutputFormat.class.getName());
sd.getSerdeInfo()
.setSerializationLib(parquet.hive.serde.ParquetHiveSerDe.class.getName());
} else {
throw new UnsupportedException(
tableDesc.getMeta().getStoreType() + " in HivecatalogStore");
}
}
sd.setSortCols(new ArrayList<Order>());
table.setSd(sd);
client.getHiveClient().createTable(table);
} catch (Throwable t) {
throw new TajoInternalError(t);
} finally {
if (client != null) client.release();
}
}
@Test
public void testHiveLocalMetaStore() {
// Create a table and display it back
try {
HiveMetaStoreClient hiveClient = new HiveMetaStoreClient(hiveLocalMetaStore.getHiveConf());
hiveClient.dropTable(
propertyParser.getProperty(ConfigVars.HIVE_TEST_DATABASE_NAME_KEY),
propertyParser.getProperty(ConfigVars.HIVE_TEST_TABLE_NAME_KEY),
true,
true);
// Define the cols
List<FieldSchema> cols = new ArrayList<FieldSchema>();
cols.add(new FieldSchema("id", serdeConstants.INT_TYPE_NAME, ""));
cols.add(new FieldSchema("msg", serdeConstants.STRING_TYPE_NAME, ""));
// Values for the StorageDescriptor
String location =
new File(propertyParser.getProperty(ConfigVars.HIVE_TEST_TABLE_NAME_KEY))
.getAbsolutePath();
String inputFormat = "org.apache.hadoop.hive.ql.io.orc.OrcInputFormat";
String outputFormat = "org.apache.hadoop.hive.ql.io.orc.OrcOutputFormat";
int numBuckets = 16;
Map<String, String> orcProps = new HashMap<String, String>();
orcProps.put("orc.compress", "NONE");
SerDeInfo serDeInfo =
new SerDeInfo(OrcSerde.class.getSimpleName(), OrcSerde.class.getName(), orcProps);
List<String> bucketCols = new ArrayList<String>();
bucketCols.add("id");
// Build the StorageDescriptor
StorageDescriptor sd = new StorageDescriptor();
sd.setCols(cols);
sd.setLocation(location);
sd.setInputFormat(inputFormat);
sd.setOutputFormat(outputFormat);
sd.setNumBuckets(numBuckets);
sd.setSerdeInfo(serDeInfo);
sd.setBucketCols(bucketCols);
sd.setSortCols(new ArrayList<Order>());
sd.setParameters(new HashMap<String, String>());
// Define the table
Table tbl = new Table();
tbl.setDbName(propertyParser.getProperty(ConfigVars.HIVE_TEST_DATABASE_NAME_KEY));
tbl.setTableName(propertyParser.getProperty(ConfigVars.HIVE_TEST_TABLE_NAME_KEY));
tbl.setSd(sd);
tbl.setOwner(System.getProperty("user.name"));
tbl.setParameters(new HashMap<String, String>());
tbl.setViewOriginalText("");
tbl.setViewExpandedText("");
tbl.setTableType(TableType.EXTERNAL_TABLE.name());
List<FieldSchema> partitions = new ArrayList<FieldSchema>();
partitions.add(new FieldSchema("dt", serdeConstants.STRING_TYPE_NAME, ""));
tbl.setPartitionKeys(partitions);
// Create the table
hiveClient.createTable(tbl);
// Describe the table
Table createdTable =
hiveClient.getTable(
propertyParser.getProperty(ConfigVars.HIVE_TEST_DATABASE_NAME_KEY),
propertyParser.getProperty(ConfigVars.HIVE_TEST_TABLE_NAME_KEY));
LOG.info("HIVE: Created Table: {}", createdTable.toString());
assertThat(
createdTable.toString(),
containsString(propertyParser.getProperty(ConfigVars.HIVE_TEST_TABLE_NAME_KEY)));
} catch (MetaException e) {
e.printStackTrace();
} catch (TException e) {
e.printStackTrace();
}
}