def test_sql_statement(self) -> None:
"""
Test Extraction with empty result from query
"""
with patch.object(SQLAlchemyExtractor, '_get_connection'):
extractor = AthenaMetadataExtractor()
extractor.init(self.conf)
self.assertTrue(self.where_clause_suffix in extractor.sql_stmt)
def test_extraction_with_empty_query_result(self) -> None:
"""
Test Extraction with empty result from query
"""
with patch.object(SQLAlchemyExtractor, '_get_connection'):
extractor = AthenaMetadataExtractor()
extractor.init(self.conf)
results = extractor.extract()
self.assertEqual(results, None)
def test_extraction_with_single_result(self):
# type: () -> None
with patch.object(SQLAlchemyExtractor, '_get_connection') as mock_connection:
connection = MagicMock()
mock_connection.return_value = connection
sql_execute = MagicMock()
connection.execute = sql_execute
table = {'schema_name': 'test_schema',
'name': 'test_table',
'description': '',
'cluster': self.conf['extractor.athena_metadata.{}'.format(AthenaMetadataExtractor.CATALOG_KEY)],
}
sql_execute.return_value = [
self._union(
{'col_name': 'col_id1',
'col_type': 'bigint',
'col_description': 'description of id1',
'col_sort_order': 0,
'extras': None}, table),
self._union(
{'col_name': 'col_id2',
'col_type': 'bigint',
'col_description': 'description of id2',
'col_sort_order': 1,
'extras': None}, table),
self._union(
{'col_name': 'is_active',
'col_type': 'boolean',
'col_description': None,
'col_sort_order': 2,
'extras': None}, table),
self._union(
{'col_name': 'source',
'col_type': 'varchar',
'col_description': 'description of source',
'col_sort_order': 3,
'extras': None}, table),
self._union(
{'col_name': 'etl_created_at',
'col_type': 'timestamp',
'col_description': None,
'col_sort_order': 4,
'extras': 'partition key'}, table),
self._union(
{'col_name': 'ds',
'col_type': 'varchar',
'col_description': None,
'col_sort_order': 5,
'extras': None}, table)
]
extractor = AthenaMetadataExtractor()
extractor.init(self.conf)
actual = extractor.extract()
expected = TableMetadata('athena', self.conf['extractor.athena_metadata.{}'.
format(AthenaMetadataExtractor.CATALOG_KEY)], 'test_schema', 'test_table', '',
[ColumnMetadata('col_id1', 'description of id1', 'bigint', 0),
ColumnMetadata('col_id2', 'description of id2', 'bigint', 1),
ColumnMetadata('is_active', None, 'boolean', 2),
ColumnMetadata('source', 'description of source', 'varchar', 3),
ColumnMetadata('etl_created_at', 'partition key', 'timestamp', 4),
ColumnMetadata('ds', None, 'varchar', 5)])
self.assertEqual(expected.__repr__(), actual.__repr__())
self.assertIsNone(extractor.extract())
