Support postgres multiple DB ingestion for user with no access (#10207)

bootstrap/sql/com.mysql.cj.jdbc.Driver/v009__create_db_connection_info.sql

@@ -5,4 +5,23 @@ ADD UNIQUE (email);
 
 -- Remove classificationName in BigQuery
 UPDATE dbservice_entity
-SET json = JSON_REMOVE(json, '$.connection.config.classificationName') where serviceType in ('BigQuery');
+SET json = JSON_REMOVE(json, '$.connection.config.classificationName') where serviceType in ('BigQuery');
+
+-- migrate ingestAllDatabases in postgres 
+UPDATE dbservice_entity de2 
+SET json = JSON_REPLACE(
+    JSON_INSERT(json, 
+      '$.connection.config.database', 
+      (select JSON_EXTRACT(json, '$.name') 
+        from database_entity de 
+        where id = (select er.toId 
+            from entity_relationship er 
+            where er.fromId = de2.id 
+              and er.toEntity = 'database' 
+            LIMIT 1
+          ))
+    ), '$.connection.config.ingestAllDatabases', 
+    true
+  ) 
+where de2.serviceType = 'Postgres' 
+  and JSON_EXTRACT(json, '$.connection.config.database') is NULL

bootstrap/sql/org.postgresql.Driver/v009__create_db_connection_info.sql

@@ -4,4 +4,24 @@ ADD UNIQUE (email);
 
 
 -- Remove classificationName in BigQuery
-UPDATE dbservice_entity SET json = json #- '{connection,config,classificationName}' where serviceType in ('BigQuery');
+UPDATE dbservice_entity SET json = json #- '{connection,config,classificationName}' where serviceType in ('BigQuery');
+
+-- migrate ingestAllDatabases in postgres 
+UPDATE dbservice_entity de2 
+SET json = JSONB_SET(
+    json || JSONB_SET(json,'{connection,config}', json#>'{connection,config}'|| 
+    jsonb_build_object('database',
+    (SELECT json->>'name' 
+        FROM database_entity de 
+        WHERE id = (SELECT er.toId 
+                    FROM entity_relationship er 
+                    WHERE er.fromId = de2.id 
+                    AND er.toEntity = 'database' 
+                    LIMIT 1)
+    )
+    )), 
+    '{connection,config,ingestAllDatabases}', 
+    'true'::jsonb
+)
+WHERE de2.serviceType = 'Postgres' 
+AND json->>'{connection,config,database}' IS NULL;

ingestion/src/metadata/ingestion/source/database/postgres/metadata.py

@@ -48,6 +48,7 @@ from metadata.ingestion.source.database.common_db_source import (
 )
 from metadata.ingestion.source.database.postgres.queries import (
     POSTGRES_GET_ALL_TABLE_PG_POLICY,
+    POSTGRES_GET_DB_NAMES,
     POSTGRES_GET_TABLE_NAMES,
     POSTGRES_PARTITION_DETAILS,
     POSTGRES_TABLE_COMMENTS,
@@ -174,14 +175,12 @@ class PostgresSource(CommonDbSourceService):
         ]
 
     def get_database_names(self) -> Iterable[str]:
-        configured_db = self.config.serviceConnection.__root__.config.database
-        if configured_db:
+        if not self.config.serviceConnection.__root__.config.ingestAllDatabases:
+            configured_db = self.config.serviceConnection.__root__.config.database
             self.set_inspector(database_name=configured_db)
             yield configured_db
         else:
-            results = self.connection.execute(
-                "select datname from pg_catalog.pg_database"
-            )
+            results = self.connection.execute(POSTGRES_GET_DB_NAMES)
             for res in results:
                 row = list(res)
                 new_database = row[0]

ingestion/src/metadata/ingestion/source/database/postgres/queries.py

@@ -132,3 +132,8 @@ POSTGRES_SQL_STATEMENT_TEST = """
         JOIN pg_catalog.pg_database d ON s.dbid = d.oid
         JOIN pg_catalog.pg_user u ON s.userid = u.usesysid
     """
+
+
+POSTGRES_GET_DB_NAMES = """
+select datname from pg_catalog.pg_database
+"""

ingestion/tests/unit/test_source_connection.py

@@ -557,16 +557,6 @@ class SourceConnectionTest(TestCase):
         assert expected_url == get_connection_url_common(mariadb_conn_obj)
 
     def test_postgres_url(self):
-        # connection arguments without db
-        expected_url = "postgresql+psycopg2://openmetadata_user:@localhost:5432"
-        postgres_conn_obj = PostgresConnection(
-            username="openmetadata_user",
-            hostPort="localhost:5432",
-            scheme=PostgresScheme.postgresql_psycopg2,
-            database=None,
-        )
-        assert expected_url == get_connection_url_common(postgres_conn_obj)
-
         # connection arguments with db
         expected_url = "postgresql+psycopg2://openmetadata_user:@localhost:5432/default"
         postgres_conn_obj = PostgresConnection(
@@ -745,7 +735,7 @@ class SourceConnectionTest(TestCase):
             username="user",
             password=None,
             hostPort="localhost:443",
-            database=None,
+            database="postgres",
             connectionArguments=None,
             scheme=PostgresScheme.postgresql_psycopg2,
         )

openmetadata-spec/src/main/resources/json/schema/entity/services/connections/database/postgresConnection.json

@@ -59,6 +59,12 @@
       "type": "string",
       "default": "PostgresPolicyTags"
     },
+    "ingestAllDatabases": {
+      "title": "Ingest All Databases",
+      "description": "Ingest data from all databases in Postgres. You can use databaseFilterPattern on top of this.",
+      "type": "boolean",
+      "default": false
+    },
     "connectionOptions": {
       "title": "Connection Options",
       "$ref": "../connectionBasicType.json#/definitions/connectionOptions"
@@ -94,5 +100,5 @@
     }
   },
   "additionalProperties": false,
-  "required": ["hostPort", "username"]
+  "required": ["hostPort", "username", "database"]
 }