Fake Snowflake Connector for Python. Run and mock Snowflake DB locally.

pip install fakesnow

Run script.py with fakesnow:

fakesnow script.py

Or a module, eg: pytest

fakesnow -m pytest

fakesnow executes fakesnow.patch before running the script or module.

To use fakesnow within your code:

import fakesnow
import snowflake.connector

with fakesnow.patch():
    conn = snowflake.connector.connect()

    print(conn.cursor().execute("SELECT 'Hello fake world!'").fetchone())

The following imports are automatically patched:

• import snowflake.connector.connect
• import snowflake.connector.pandas_tools.write_pandas

To patch modules that use the from ... import syntax, manually specify them, eg: if mymodule.py has the import:

from snowflake.connector.pandas_tools import write_pandas

Then patch it using:

with fakesnow.patch("mymodule.write_pandas"):
    ...

By default databases are in-memory. To persist databases between processes, specify a databases path:

with fakesnow.patch(db_path="databases/"):
    ...

pytest fixtures are provided for testing. Example conftest.py:

from typing import Iterator

import fakesnow.fixtures
import pytest

pytest_plugins = fakesnow.fixtures.__name__

@pytest.fixture(scope="session", autouse=True)
def setup(_fakesnow_session: None) -> Iterator[None]:
    # the standard imports are now patched
    ...
    yield

Or with from ... import patch targets:

from typing import Iterator

import fakesnow
import pytest

@pytest.fixture(scope="session", autouse=True)
def _fakesnow_session() -> Iterator[None]:
    with fakesnow.patch("mymodule.write_pandas"):
        yield

• cursors and standard SQL
• get_result_batches()
• information schema
• multiple databases
• parameter binding
• table comments
• write_pandas(..)
• access control
• standalone/out of process api/support for faking non-python connectors
• stored procedures

Partial support

• date functions
• regex functions
• semi-structured data
• tags
• user management (See tests/test_users.py)

For more detail see tests/test_fakes.py

• The order of rows is non deterministic and may not match Snowflake unless ORDER BY is fully specified.
• A more liberal Snowflake SQL dialect than used by a real Snowflake instance is supported, ie: some queries might pass using fakesnow that a real Snowflake instance would reject.

See CONTRIBUTING.md to get started and develop in this repo.