ADBC Snowflake Driver

1.11.0 verified Fri May 01 auth: no python

An ADBC (Arrow Database Connectivity) driver for Snowflake, enabling high-performance data transfer using Apache Arrow. Version 1.11.0 requires Python >=3.10.

pip install adbc-driver-snowflake

Common errors

error ImportError: cannot import name 'connect' from 'adbc_driver_snowflake' ↓

cause Trying to import connect from the top-level module instead of submodule dbapi.

fix

Use: from adbc_driver_snowflake import dbapi; conn = dbapi.connect(...)

error Exception: No suitable driver found ↓

cause Missing ADBC driver manager or incorrect installation of the Snowflake driver.

fix

pip install adbc-driver-manager adbc-driver-snowflake

error pyarrow.lib.ArrowNotImplementedError: Unsupported cast from string to int64 using function cast ↓

cause Incompatible schema or types between Snowflake and Arrow; often due to NULL handling.

fix

Ensure Snowflake columns are typed appropriately; use CAST in SQL if needed.

Warnings

breaking The driver is not a DB-API 2.0 module directly; you must use the .dbapi submodule. Direct import of adbc_driver_snowflake does not expose connect() ↓

fix Use from adbc_driver_snowflake import dbapi, then dbapi.connect()

gotcha Connection string must include warehouse as a query parameter; omitted warehouse may lead to errors about virtual warehouse not specified. ↓

fix Append ?warehouse=my_wh to the Snowflake URL.

gotcha The driver only supports Arrow data; fetching results via .fetchall() returns a list of Arrow arrays, not plain Python objects. Use fetch_arrow_batch() to get a RecordBatch. ↓

fix Use .fetch_arrow_batch() or .fetch_arrow_table() for pandas conversion.

Imports

adbc_driver_snowflake
wrong
```
import adbc_driver_snowflake.dbapi
```
correct
```
import adbc_driver_snowflake
```
Wrong submodule path; top-level import provides dbapi via adbc_driver_snowflake.dbapi

Quickstart

Connect to Snowflake using ADBC and fetch an Arrow batch converted to pandas.

import adbc_driver_snowflake.dbapi
conn = adbc_driver_snowflake.dbapi.connect(
    'snowflake://user:pass@account/database/schema?warehouse=warehouse_name'
)
cur = conn.cursor()
cur.execute("SELECT 1 AS col")
print(cur.fetch_arrow_batch().to_pandas())