make discovery work when user has SELECT on only some columns on a table

tap_postgres/__init__.py

@@ -273,7 +273,7 @@ def produce_table_info(conn):
 AND NOT a.attisdropped
 AND pg_class.relkind IN ('r', 'v', 'm')
 AND n.nspname NOT in ('pg_toast', 'pg_catalog', 'information_schema')
-AND has_table_privilege(pg_class.oid, 'SELECT') = true """)
+AND has_column_privilege(pg_class.oid, attname, 'SELECT') = true """)
         for row in cur.fetchall():
             row_count, is_view, schema_name, table_name, *col_info = row
 

tests/test_discovery.py

@@ -576,6 +576,63 @@ def test_catalog(self):
                                   'definitions' : tap_postgres.BASE_RECURSIVE_SCHEMAS},
                                  stream_dict.get('schema'))
 
+class TestColumnGrants(unittest.TestCase):
+    maxDiff = None
+    table_name = 'CHICKEN TIMES'
+    user = 'tmp_user_for_grant_tests'
+    password = 'password'
+
+    def setUp(self):
+        table_spec = {"columns": [{"name" : "id",                "type" : "integer",  "serial" : True},
+                                 {"name" : 'size integer',      "type" : "integer",  "quoted" : True},
+                                 {"name" : 'size smallint',     "type" : "smallint", "quoted" : True},
+                                 {"name" : 'size bigint',       "type" : "bigint",   "quoted" : True}],
+                     "name" : TestColumnGrants.table_name}
+        ensure_test_table(table_spec)
+
+        with get_test_connection() as conn:
+           cur = conn.cursor()
+
+           sql = """ DROP USER IF EXISTS {} """.format(self.user, self.password)
+           LOGGER.info(sql)
+           cur.execute(sql)
+
+           sql = """ CREATE USER {} WITH PASSWORD '{}' """.format(self.user, self.password)
+           LOGGER.info(sql)
+           cur.execute(sql)
+
+           sql = """ GRANT SELECT ("id") ON "{}" TO {}""".format(TestColumnGrants.table_name, self.user)
+           LOGGER.info("running sql: {}".format(sql))
+           cur.execute(sql)
+
+
+
+
+    def test_catalog(self):
+        conn_config = get_test_connection_config()
+        conn_config['user'] = self.user
+        conn_config['password'] = self.password
+        streams = tap_postgres.do_discovery(conn_config)
+        chicken_streams = [s for s in streams if s['tap_stream_id'] == 'postgres-public-CHICKEN TIMES']
+
+        self.assertEqual(len(chicken_streams), 1)
+        stream_dict = chicken_streams[0]
+
+        self.assertEqual(TestStringTableWithPK.table_name, stream_dict.get('table_name'))
+        self.assertEqual(TestStringTableWithPK.table_name, stream_dict.get('stream'))
+
+
+        stream_dict.get('metadata').sort(key=lambda md: md['breadcrumb'])
+
+        self.assertEqual(metadata.to_map(stream_dict.get('metadata')),
+                         {() : {'table-key-properties': [], 'database-name': 'postgres', 'schema-name': 'public', 'is-view': False, 'row-count': 0},
+                          ('properties', 'id')                     : {'inclusion': 'available', 'sql-datatype' : 'integer', 'selected-by-default' : True}})
+
+        self.assertEqual({'definitions' : tap_postgres.BASE_RECURSIVE_SCHEMAS,
+                          'type': 'object',
+                          'properties': {'id': {'type': ['null', 'integer'], 'minimum': -2147483648, 'maximum': 2147483647}}},
+                         stream_dict.get('schema'))
+
 
 if __name__== "__main__":
     test1 = TestArraysTable()