Using fetch_pandas_all() or fetch_arrow_all() without installing pyarrow. pyarrow was uninstalled or excluded from the dependency list during a package update. Using a virtual environment that does not have pyarrow installed. Deploying to a container image that does not include pyarrow in requirements.txt. Using write_pandas() from snowflake.connector.pandas_tools without pyarrow

Install pyarrow: pip install pyarrow.. For connector compatibility, install the version specified in Snowflake's connector extras: pip install snowflake-connector-python[pandas].. Verify the install is in the correct environment: which python and pip list | grep pyarrow.. Add pyarrow to requirements.txt or pyproject.toml and rebuild the container/env.. If pyarrow is not available and cannot be installed, use cursor.fetchall() and convert to a DataFrame manually.

Low severityconfigurationSnowflake →

Power BI Refresh Error:
255002

What does this error mean?

The Python connector attempted to use Arrow-based result fetching but the pyarrow package is not installed. Arrow format is used for fetch_pandas_all(), fetch_arrow_all(), and the write_pandas() helper — all of which require pyarrow.

Common causes

1Using fetch_pandas_all() or fetch_arrow_all() without installing pyarrow
2pyarrow was uninstalled or excluded from the dependency list during a package update
3Using a virtual environment that does not have pyarrow installed
4Deploying to a container image that does not include pyarrow in requirements.txt
5Using write_pandas() from snowflake.connector.pandas_tools without pyarrow

How to fix it

1Install pyarrow: pip install pyarrow.
2For connector compatibility, install the version specified in Snowflake's connector extras: pip install snowflake-connector-python[pandas].
3Verify the install is in the correct environment: which python and pip list | grep pyarrow.
4Add pyarrow to requirements.txt or pyproject.toml and rebuild the container/env.
5If pyarrow is not available and cannot be installed, use cursor.fetchall() and convert to a DataFrame manually.

Frequently asked questions

Do I need pyarrow for basic Snowflake queries?

No — pyarrow is only needed for Arrow-based fetching (fetch_pandas_all, fetch_arrow_all, write_pandas). fetchall() and fetchone() work without it.

Which pyarrow version is compatible with my connector version?

Install with pip install 'snowflake-connector-python[pandas]' — this resolves the correct pyarrow version automatically based on the connector version.

Official documentation: https://github.com/snowflakedb/snowflake-connector-python/blob/main/src/snowflake/connector/errorcode.py