def test_empty_connection(self):
"""Check the error when instantiating with an empty
connection string"""
pattern = (r"`con` argument cannot be empty\. "
r"Please provide a SQLAlchemy connection string\.")
with pytest.raises(DataSetError, match=pattern):
SQLTableDataSet(table_name=TABLE_NAME, credentials=dict(con=""))
def test_create_connection_only_once(self, mocker):
"""Test that two datasets that need to connect to the same db
(but different tables, for example) only create a connection once.
"""
mock_engine = mocker.patch(
"kedro.extras.datasets.pandas.sql_dataset.create_engine")
first = SQLTableDataSet(table_name=TABLE_NAME,
credentials=dict(con=CONNECTION))
assert len(first.engines) == 1
second = SQLTableDataSet(table_name="other_table",
credentials=dict(con=CONNECTION))
assert len(second.engines) == 1
assert len(first.engines) == 1
mock_engine.assert_called_once_with(CONNECTION)
def test_unknown_sql(self):
"""Check the error when unknown sql dialect is provided;
this means the error is raised on catalog creation, rather
than on load or save operation.
"""
pattern = r"The SQL dialect in your connection is not supported by SQLAlchemy"
with pytest.raises(DataSetError, match=pattern):
SQLTableDataSet(table_name=TABLE_NAME,
credentials=dict(con=FAKE_CONN_STR))
def test_driver_missing(self, mocker):
"""Check the error when the sql driver is missing"""
mocker.patch(
"kedro.extras.datasets.pandas.sql_dataset.create_engine",
side_effect=ImportError("No module named 'mysqldb'"),
)
with pytest.raises(DataSetError, match=ERROR_PREFIX + "mysqlclient"):
SQLTableDataSet(table_name=TABLE_NAME,
credentials=dict(con=CONNECTION))
def test_multiple_connections(self, mocker):
"""Test that two datasets that need to connect to different dbs
only create one connection per db.
"""
mock_engine = mocker.patch(
"kedro.extras.datasets.pandas.sql_dataset.create_engine")
first = SQLTableDataSet(table_name=TABLE_NAME,
credentials=dict(con=CONNECTION))
assert len(first.engines) == 1
second_con = f"other_{CONNECTION}"
second = SQLTableDataSet(table_name=TABLE_NAME,
credentials=dict(con=second_con))
assert len(second.engines) == 2
assert len(first.engines) == 2
expected_calls = [mocker.call(CONNECTION), mocker.call(second_con)]
assert mock_engine.call_args_list == expected_calls
def test_unknown_module(self, mocker):
"""Test that if an unknown module/driver is encountered by SQLAlchemy
then the error should contain the original error message"""
mocker.patch(
"kedro.extras.datasets.pandas.sql_dataset.create_engine",
side_effect=ImportError("No module named 'unknown_module'"),
)
pattern = ERROR_PREFIX + r"No module named \'unknown\_module\'"
with pytest.raises(DataSetError, match=pattern):
SQLTableDataSet(table_name=TABLE_NAME,
credentials=dict(con=CONNECTION))
def test_single_connection(self, dummy_dataframe, mocker):
"""Test to make sure multiple instances use the same connection object."""
mocker.patch("pandas.read_sql_table")
dummy_to_sql = mocker.patch.object(dummy_dataframe, "to_sql")
kwargs = dict(table_name=TABLE_NAME, credentials=dict(con=CONNECTION))
first = SQLTableDataSet(**kwargs)
unique_connection = first.engines[CONNECTION]
datasets = [SQLTableDataSet(**kwargs) for _ in range(10)]
for ds in datasets:
ds.save(dummy_dataframe)
engine = ds.engines[CONNECTION]
assert engine is unique_connection
expected_call = mocker.call(name=TABLE_NAME,
con=unique_connection,
index=False)
dummy_to_sql.assert_has_calls([expected_call] * 10)
for ds in datasets:
ds.load()
engine = ds.engines[CONNECTION]
assert engine is unique_connection
def test_empty_table_name(self):
"""Check the error when instantiating with an empty table"""
pattern = r"`table\_name` argument cannot be empty\."
with pytest.raises(DataSetError, match=pattern):
SQLTableDataSet(table_name="", credentials=dict(con=CONNECTION))
def table_data_set(request):
kwargs = dict(table_name=TABLE_NAME, credentials=dict(con=CONNECTION))
kwargs.update(request.param)
return SQLTableDataSet(**kwargs)
#Read New Matched SCMH Data
"read_new_smch_matched": SQLQueryDataSet(
sql= read_new_smch_matched,
credentials=dict(con=con)
),
#Read Old Matched SCMH Data
"read_old_smch_matched_data": SQLQueryDataSet(
sql= read_old_smch_matched_data,
credentials=dict(con=con)
),
#Make Use Of Save Method To Create Tables
"create_derived_admissions": SQLTableDataSet(
table_name='admissions',
credentials=dict(con=con),
save_args = dict(schema='derived',if_exists='replace')
),
#Make Use Of Save Method To Create Tables
"create_derived_discharges": SQLTableDataSet(
table_name="discharges",
credentials=dict(con=con),
save_args = dict(schema="derived",if_exists="replace")
),
#Make Use Of Save Method To Create Tables
"create_joined_admissions_discharges": SQLTableDataSet(
table_name='joined_admissions_discharges',
credentials=dict(con=con),
save_args = dict(schema='derived',if_exists='replace')
),
