JonasJ-ap commented on code in PR #6034: URL: https://github.com/apache/iceberg/pull/6034#discussion_r1008989781
########## python/pyiceberg/catalog/glue.py: ########## @@ -0,0 +1,453 @@ +# Licensed to the Apache Software Foundation (ASF) under one +# or more contributor license agreements. See the NOTICE file +# distributed with this work for additional information +# regarding copyright ownership. The ASF licenses this file +# to you under the Apache License, Version 2.0 (the +# "License"); you may not use this file except in compliance +# with the License. You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, +# software distributed under the License is distributed on an +# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY +# KIND, either express or implied. See the License for the +# specific language governing permissions and limitations +# under the License. + + +import uuid +from typing import ( + Any, + Dict, + List, + Optional, + Set, + Union, +) + +import boto3 + +from pyiceberg.catalog import ( + Catalog, + Identifier, + Properties, + PropertiesUpdateSummary, +) +from pyiceberg.catalog.base import ( + ICEBERG, + METADATA_LOCATION, + TABLE_TYPE, + WAREHOUSE, +) +from pyiceberg.exceptions import ( + NamespaceAlreadyExistsError, + NamespaceNotEmptyError, + NoSuchIcebergTableError, + NoSuchNamespaceError, + NoSuchPropertyException, + NoSuchTableError, + TableAlreadyExistsError, +) +from pyiceberg.io import FileIO, load_file_io +from pyiceberg.schema import Schema +from pyiceberg.serializers import FromInputFile, ToOutputFile +from pyiceberg.table import Table +from pyiceberg.table.metadata import TableMetadata, new_table_metadata +from pyiceberg.table.partitioning import UNPARTITIONED_PARTITION_SPEC, PartitionSpec +from pyiceberg.table.sorting import UNSORTED_SORT_ORDER, SortOrder +from pyiceberg.typedef import EMPTY_DICT + +EXTERNAL_TABLE_TYPE = "EXTERNAL_TABLE" + +PROP_GLUE_TABLE = "Table" +PROP_GLUE_TABLE_TYPE = "TableType" +PROP_GLUE_TABLE_DESCRIPTION = "description" +PROP_GLUE_TABLE_PARAMETERS = "Parameters" +PROP_GLUE_TABLE_DATABASE_NAME = "DatabaseName" +PROP_GLUE_TABLE_NAME = "Name" + +PROP_GLUE_TABLELIST = "TableList" + +PROP_GLUE_DATABASE = "Database" +PROP_GLUE_DATABASE_LIST = "DatabaseList" +PROP_GLUE_DATABASE_NAME = "Name" +PROP_GLUE_DATABASE_LOCATION = "LocationUri" +PROP_GLUE_DATABASE_DESCRIPTION = "Description" +PROP_GLUE_DATABASE_PARAMETERS = "Parameters" + +PROP_GLUE_NEXT_TOKEN = "NextToken" + +GLUE_DESCRIPTION_KEY = "comment" +GLUE_DATABASE_LOCATION_KEY = "location" + + +def _construct_parameters(metadata_location: str) -> Properties: + return {TABLE_TYPE: ICEBERG.upper(), METADATA_LOCATION: metadata_location} + + +def _construct_table_input(table_name: str, metadata_location: str, properties: Properties) -> Dict[str, Any]: + table_input = { + PROP_GLUE_TABLE_NAME: table_name, + PROP_GLUE_TABLE_TYPE: EXTERNAL_TABLE_TYPE, + PROP_GLUE_TABLE_PARAMETERS: _construct_parameters(metadata_location), + } + + if table_description := properties.get(GLUE_DESCRIPTION_KEY): + table_input[PROP_GLUE_TABLE_DESCRIPTION] = table_description + + return table_input + + +def _construct_database_input(database_name: str, properties: Properties) -> Dict[str, Any]: + database_input: Dict[str, Any] = {PROP_GLUE_DATABASE_NAME: database_name} + parameters = {} + for k, v in properties.items(): + if k == GLUE_DESCRIPTION_KEY: + database_input[PROP_GLUE_DATABASE_DESCRIPTION] = v + if k == GLUE_DATABASE_LOCATION_KEY: + database_input[PROP_GLUE_DATABASE_LOCATION] = v + else: + parameters[k] = v + database_input[PROP_GLUE_DATABASE_PARAMETERS] = parameters + return database_input + + +def _write_metadata(metadata: TableMetadata, io: FileIO, metadate_path: str): + ToOutputFile.table_metadata(metadata, io.new_output(metadate_path)) + + +class GlueCatalog(Catalog): + def __init__(self, name: str, **properties: str): + super().__init__(name, **properties) + self.glue = boto3.client("glue") + + def _convert_glue_to_iceberg(self, glue_table: Dict[str, Any]) -> Table: + properties: Properties = glue_table[PROP_GLUE_TABLE_PARAMETERS] + + if TABLE_TYPE not in properties: + raise NoSuchPropertyException( + f"Property {TABLE_TYPE} missing, could not determine type: " + f"{glue_table[PROP_GLUE_TABLE_DATABASE_NAME]}.{glue_table[PROP_GLUE_TABLE_NAME]}" + ) + glue_table_type = properties[TABLE_TYPE] + + if glue_table_type.lower() != ICEBERG: + raise NoSuchIcebergTableError( + f"Property table_type is {glue_table_type}, expected {ICEBERG}: " + f"{glue_table[PROP_GLUE_TABLE_DATABASE_NAME]}.{glue_table[PROP_GLUE_TABLE_NAME]}" + ) + + if METADATA_LOCATION not in properties: + raise NoSuchPropertyException( + f"Table property {METADATA_LOCATION} is missing, cannot find metadata for: " + f"{glue_table[PROP_GLUE_TABLE_DATABASE_NAME]}.{glue_table[PROP_GLUE_TABLE_NAME]}" + ) + metadata_location = properties[METADATA_LOCATION] + + io = load_file_io(properties=self.properties, location=metadata_location) + file = io.new_input(metadata_location) + metadata = FromInputFile.table_metadata(file) + return Table( + identifier=(glue_table[PROP_GLUE_TABLE_DATABASE_NAME], glue_table[PROP_GLUE_TABLE_NAME]), + metadata=metadata, + metadata_location=metadata_location, + ) + + def _default_warehouse_location(self, database_name: str, table_name: str): + try: + response = self.glue.get_database(Name=database_name) + except self.glue.exceptions.EntityNotFoundException as e: + raise NoSuchNamespaceError(f"The database: {database_name} does not exist") from e + + if database_location := response.get(PROP_GLUE_DATABASE).get(PROP_GLUE_DATABASE_LOCATION): + return f"{database_location}/{table_name}" + + if WAREHOUSE in self.properties: + return f"{self.properties[WAREHOUSE]}/{database_name}.db/{table_name}" + + raise ValueError("No default path is set, please specify a location when creating a table") + + def _resolve_table_location(self, location: Optional[str], database_name: str, table_name: str) -> str: + if not location: + return self._default_warehouse_location(database_name, table_name) + return location + + def create_table( + self, + identifier: Union[str, Identifier], + schema: Schema, + location: Optional[str] = None, + partition_spec: PartitionSpec = UNPARTITIONED_PARTITION_SPEC, + sort_order: SortOrder = UNSORTED_SORT_ORDER, + properties: Properties = EMPTY_DICT, + ) -> Table: + """Create an Iceberg table in Glue catalog + Args: + identifier: Table identifier. + schema: Table's schema. + location: Location for the table. Optional Argument. + partition_spec: PartitionSpec for the table. + sort_order: SortOrder for the table. + properties: Table properties that can be a string based dictionary. + Returns: + Table: the created table instance + Raises: + AlreadyExistsError: If a table with the name already exists + ValueError: If the identifier is invalid + """ + database_name, table_name = self.identifier_to_tuple(identifier) + + location = self._resolve_table_location(location, database_name, table_name) + metadata_location = f"{location}/metadata/00000-{uuid.uuid4()}.metadata.json" + metadata = new_table_metadata( + location=location, schema=schema, partition_spec=partition_spec, sort_order=sort_order, properties=properties + ) + io = load_file_io(properties=self.properties, location=metadata_location) + _write_metadata(metadata, io, metadata_location) + try: + self.glue.create_table( + DatabaseName=database_name, TableInput=_construct_table_input(table_name, metadata_location, properties) + ) + except self.glue.exceptions.AlreadyExistsException as e: + raise TableAlreadyExistsError(f"Table {database_name}.{table_name} already exists") from e + except self.glue.exceptions.EntityNotFoundException as e: + raise NoSuchNamespaceError(f"Database {database_name} not found") from e + + try: + loaded_table = self.load_table(identifier=(database_name, table_name)) + except self.glue.exceptions.EntityNotFoundException as e: + raise NoSuchTableError(f"Table {database_name}.{table_name} fail to be created") from e + + return loaded_table + + def load_table(self, identifier: Union[str, Identifier]) -> Table: + """Loads the table's metadata and returns the table instance. + You can also use this method to check for table existence using 'try catalog.table() except TableNotFoundError' + Note: This method doesn't scan data stored in the table. + Args: + identifier: Table identifier. + Returns: + Table: the table instance with its metadata + Raises: + NoSuchTableError: If a table with the name does not exist, or the identifier is invalid + """ + database_name, table_name = self.identifier_to_tuple(identifier) + try: + load_table_response = self.glue.get_table(DatabaseName=database_name, Name=table_name) + except self.glue.exceptions.EntityNotFoundException as e: + raise NoSuchTableError(f"Table does not exists: {database_name}.{table_name}") from e + + return self._convert_glue_to_iceberg(load_table_response[PROP_GLUE_TABLE]) + + def drop_table(self, identifier: Union[str, Identifier]) -> None: + """Drop a table. + + Args: + identifier: Table identifier. + + Raises: + NoSuchTableError: If a table with the name does not exist, or the identifier is invalid + """ + database_name, table_name = self.identifier_to_tuple(identifier) + try: + self.glue.delete_table(DatabaseName=database_name, Name=table_name) + except self.glue.exceptions.EntityNotFoundException as e: + raise NoSuchTableError(f"Table does not exists: {database_name}.{table_name}") from e + + def purge_table(self, identifier: Union[str, Identifier]) -> None: + """Drop a table and purge all data and metadata files. AWS glue will asynchronously delete resource + belongs to the table once the table is deleted + + Args: + identifier (str | Identifier): Table identifier. + + Raises: + NoSuchTableError: If a table with the name does not exist + """ + self.drop_table(identifier) + + def rename_table(self, from_identifier: Union[str, Identifier], to_identifier: Union[str, Identifier]) -> Table: + """Rename a fully classified table name + + Args: + from_identifier: Existing table identifier. + to_identifier: New table identifier. + + Returns: + Table: the updated table instance with its metadata + + Raises: + ValueError: When the from table identifier is invalid + NoSuchTableError: When a table with the name does not exist + NoSuchNamespaceError: When the destination namespace doesn't exists + """ + from_database_name, from_table_name = self.identifier_to_tuple(from_identifier) + to_database_name, to_table_name = self.identifier_to_tuple(to_identifier) + try: + get_table_response = self.glue.get_table(DatabaseName=from_database_name, Name=from_table_name) + except self.glue.exceptions.EntityNotFoundException as e: + raise NoSuchTableError(f"Table does not exists: {from_database_name}.{from_table_name}") from e + + table_input = get_table_response[PROP_GLUE_TABLE] + table_input[PROP_GLUE_TABLE_NAME] = to_table_name + new_input = _construct_table_input(to_table_name, table_input[PROP_GLUE_TABLE_PARAMETERS][METADATA_LOCATION], {}) + + try: + self.glue.create_table(DatabaseName=to_database_name, TableInput=new_input) + except self.glue.exceptions.AlreadyExistsException as e: + raise TableAlreadyExistsError(f"Table {to_database_name}.{to_table_name} already exists") from e + except self.glue.exceptions.EntityNotFoundException as e: + raise NoSuchNamespaceError(f"Database {to_database_name} not found") from e + + try: + self.drop_table(from_identifier) + except Exception as e: + self.drop_table(to_identifier) + raise ValueError( + f"Fail to drop old table {from_database_name}.{from_table_name}, roll back to use the old one" + ) from e + return self.load_table(to_identifier) + + def create_namespace(self, namespace: Union[str, Identifier], properties: Properties = EMPTY_DICT) -> None: + """Create a namespace in the catalog. + + Args: + namespace: Namespace identifier + properties: A string dictionary of properties for the given namespace + + Raises: + ValueError: If the identifier is invalid + AlreadyExistsError: If a namespace with the given name already exists + """ + database_name = self.identifier_to_tuple(namespace)[0] + try: + self.glue.create_database(DatabaseInput=_construct_database_input(database_name, properties)) + except self.glue.exceptions.AlreadyExistsException as e: + raise NamespaceAlreadyExistsError(f"Database {database_name} already exists") from e + + def drop_namespace(self, namespace: Union[str, Identifier]) -> None: + """Drop a namespace. + + Args: + namespace: Namespace identifier + + Raises: + NoSuchNamespaceError: If a namespace with the given name does not exist, or the identifier is invalid + NamespaceNotEmptyError: If the namespace is not empty + """ + database_name = self.identifier_to_tuple(namespace)[0] + try: + table_list = self.list_tables(namespace=database_name) + except NoSuchNamespaceError as e: + raise NoSuchNamespaceError(f"Database does not exists: {database_name}") from e + + if len(table_list) > 0: + raise NamespaceNotEmptyError(f"Database {database_name} is not empty") + + self.glue.delete_database(Name=database_name) + + def list_tables(self, namespace: Union[str, Identifier]) -> List[Identifier]: + database_name = self.identifier_to_tuple(namespace)[0] + table_list = [] + try: + table_list_response = self.glue.get_tables(DatabaseName=database_name) + next_token = table_list_response.get(PROP_GLUE_NEXT_TOKEN) + table_list += table_list_response[PROP_GLUE_TABLELIST] Review Comment: Thank you for your suggestion. I think it is a good idea to change it to `.get(PROP_GLUE_TABLELIST, [])`. At the same time, I also do some tests to verify the behavior of the API calls: According to my test:https://gist.github.com/JonasJ-ap/d75735124951fa11f2bb440cc0b5a9ed, the responses of `get_tables` and `get_databases` will contain `TableList`/`DatabaseList` field even when there is no table/database in the database/aws account. I think there will be no key error in case of empty table list even in the current implementation. Moreover, according to the Java's awssdk, the `tableList()` and `databaseList()` will never be `null`. Hence, I think for boto3's corresponding method, `tableList` and `databaseList` field should always exist similar to java impl. The only difference between java's sdk and boto3 is that java's sdk provides an additional method `hasDatabaseList()` to check the internal behavior of the service. Please correct me if I misunderstand something. If I am wrong, some additional field checks will be needed in some other parts of this implementation. Thank you in advance for your help. Relevant docs: https://sdk.amazonaws.com/java/api/latest/software/amazon/awssdk/services/glue/model/GetDatabasesResponse.html https://boto3.amazonaws.com/v1/documentation/api/latest/reference/services/glue.html#Glue.Client.get_databases -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: issues-unsubscr...@iceberg.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: issues-unsubscr...@iceberg.apache.org For additional commands, e-mail: issues-h...@iceberg.apache.org
