Source code for airflow.providers.databricks.hooks.databricks_sql

# 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 re
from contextlib import closing
from copy import copy
from typing import Any, Dict, List, Optional, Union

from databricks import sql
from databricks.sql.client import Connection

from airflow.exceptions import AirflowException
from airflow.hooks.dbapi import DbApiHook
from airflow.providers.databricks.hooks.databricks_base import BaseDatabricksHook

[docs]LIST_SQL_ENDPOINTS_ENDPOINT = ('GET', 'api/2.0/sql/endpoints')
[docs]class DatabricksSqlHook(BaseDatabricksHook, DbApiHook): """ Interact with Databricks SQL. :param databricks_conn_id: Reference to the :ref:`Databricks connection <howto/connection:databricks>`. :param http_path: Optional string specifying HTTP path of Databricks SQL Endpoint or cluster. If not specified, it should be either specified in the Databricks connection's extra parameters, or ``sql_endpoint_name`` must be specified. :param sql_endpoint_name: Optional name of Databricks SQL Endpoint. If not specified, ``http_path`` must be provided as described above. :param session_configuration: An optional dictionary of Spark session parameters. Defaults to None. If not specified, it could be specified in the Databricks connection's extra parameters. """
[docs] hook_name = 'Databricks SQL'
def __init__( self, databricks_conn_id: str = BaseDatabricksHook.default_conn_name, http_path: Optional[str] = None, sql_endpoint_name: Optional[str] = None, session_configuration: Optional[Dict[str, str]] = None, ) -> None: super().__init__(databricks_conn_id) self._sql_conn = None self._token: Optional[str] = None self._http_path = http_path self._sql_endpoint_name = sql_endpoint_name self.supports_autocommit = True self.session_config = session_configuration def _get_extra_config(self) -> Dict[str, Optional[Any]]: extra_params = copy(self.databricks_conn.extra_dejson) for arg in ['http_path', 'session_configuration'] + self.extra_parameters: if arg in extra_params: del extra_params[arg] return extra_params def _get_sql_endpoint_by_name(self, endpoint_name) -> Dict[str, Any]: result = self._do_api_call(LIST_SQL_ENDPOINTS_ENDPOINT) if 'endpoints' not in result: raise AirflowException("Can't list Databricks SQL endpoints") lst = [endpoint for endpoint in result['endpoints'] if endpoint['name'] == endpoint_name] if len(lst) == 0: raise AirflowException(f"Can't f Databricks SQL endpoint with name '{endpoint_name}'") return lst[0]
[docs] def get_conn(self) -> Connection: """Returns a Databricks SQL connection object""" if not self._http_path: if self._sql_endpoint_name: endpoint = self._get_sql_endpoint_by_name(self._sql_endpoint_name) self._http_path = endpoint['odbc_params']['path'] elif 'http_path' in self.databricks_conn.extra_dejson: self._http_path = self.databricks_conn.extra_dejson['http_path'] else: raise AirflowException( "http_path should be provided either explicitly, " "or in extra parameter of Databricks connection, " "or sql_endpoint_name should be specified" ) requires_init = True if not self._token: self._token = self._get_token(raise_error=True) else: new_token = self._get_token(raise_error=True) if new_token != self._token: self._token = new_token else: requires_init = False if not self.session_config: self.session_config = self.databricks_conn.extra_dejson.get('session_configuration') if not self._sql_conn or requires_init: if self._sql_conn: # close already existing connection self._sql_conn.close() self._sql_conn = sql.connect( self.host, self._http_path, self._token, session_configuration=self.session_config, **self._get_extra_config(), ) return self._sql_conn
@staticmethod
[docs] def maybe_split_sql_string(sql: str) -> List[str]: """ Splits strings consisting of multiple SQL expressions into an TODO: do we need something more sophisticated? :param sql: SQL string potentially consisting of multiple expressions :return: list of individual expressions """ splits = [s.strip() for s in re.split(";\\s*\r?\n", sql) if s.strip() != ""] return splits
[docs] def run(self, sql: Union[str, List[str]], autocommit=True, parameters=None, handler=None): """ Runs a command or a list of commands. Pass a list of sql statements to the sql parameter to get them to execute sequentially :param sql: the sql statement to be executed (str) or a list of sql statements to execute :param autocommit: What to set the connection's autocommit setting to before executing the query. :param parameters: The parameters to render the SQL query with. :param handler: The result handler which is called with the result of each statement. :return: query results. """ if isinstance(sql, str): sql = self.maybe_split_sql_string(sql) self.log.debug("Executing %d statements", len(sql)) conn = None for sql_statement in sql: # when using AAD tokens, it could expire if previous query run longer than token lifetime conn = self.get_conn() with closing(conn.cursor()) as cur: self.log.info("Executing statement: '%s', parameters: '%s'", sql_statement, parameters) if parameters: cur.execute(sql_statement, parameters) else: cur.execute(sql_statement) schema = cur.description results = [] if handler is not None: cur = handler(cur) for row in cur: self.log.debug("Statement results: %s", row) results.append(row) self.log.info("Rows affected: %s", cur.rowcount) if conn: conn.close() self._sql_conn = None # Return only result of the last SQL expression return schema, results
[docs] def test_connection(self): """Test the Databricks SQL connection by running a simple query.""" try: self.run(sql="select 42") except Exception as e: return False, str(e) return True, "Connection successfully checked"
[docs] def bulk_dump(self, table, tmp_file): raise NotImplementedError()
[docs] def bulk_load(self, table, tmp_file): raise NotImplementedError()

Was this entry helpful?