Devart Python Connector for Google BigQuery: DB API Access to BigQuery on Windows
Devart Python Connector for Google BigQuery is a paid Python module, $74.95 a year for one developer, that runs GoogleSQL queries against BigQuery from Python 3.10-3.15 on Windows only, through the standard Python DB API 2.0 interface. It signs in with a Google account (OAuth refresh token) or, since version 1.4, a service account JSON key, and adds connection pooling, local caching and batched updates.
Google's own free client library, google-cloud-bigquery, is the usual way to reach BigQuery from Python and runs on every platform. Devart's connector is for Windows teams that prefer a plain DB API cursor with ? parameters over Google's job-based client, or that want BigQuery under the same license and support as their other Devart connectors.
Quick facts
- Version: 1.4.0, released October 22, 2026; first released December 1, 2023
- Developer: Devart
- Package:
pip install devart-bigquery-connector, or wheel files from Devart's download page (18.21 MB for Windows) - Python: 3.10 to 3.15 (1.4 dropped 3.7 to 3.9)
- Platforms: Windows and Windows Server, 32-bit and 64-bit; no macOS or Linux build
- Sign-in: Google OAuth refresh token, or service account JSON key (added in 1.4)
- Price: $74.95 a year or $149.95 perpetual for a Single (one developer) license
- Trial: 30 days, for evaluation only, not for production use
- Refund: the order page promises a 30-day money-back guarantee "if you are not satisfied", while Devart's ordering FAQ grants refunds within 30 days only when a technical problem support can't fix stops the product working
- What we checked: the order page charged $74.95 for a Single license with one year of updates on a yearly subscription, and PyPI and Devart's download page both showed version 1.4.0 with Windows-only wheels, on November 18, 2026. We didn't run a query against a BigQuery project
Download Free Trial Buy Now
Overview
You import devart.bigquery, call bigquery.signin() to get a refresh token, and connect with bigquery.connect(ProjectId=..., DataSetId=..., RefreshToken=...). Version 1.4 also accepts a service account's JSON key, which suits scheduled jobs with no person to sign in. After that it's a cursor, execute() and fetchall(), like any DB API module.
The connector supports BigQuery's SQL dialects and data types, including aggregate, array, datetime, geography, JSON, navigation, net and hash functions, and maps BigQuery types to Python types with options to change the mapping.
Features
GoogleSQL through a DB API cursor
Python Connector for Google BigQuery runs BigQuery SQL with its functions, operators and conditional expressions through a standard cursor, with ? placeholders for parameters. Results come back as ordinary rows from fetchall(), not as job objects.
Service account sign-in
Python Connector for Google BigQuery 1.4 added service account authentication with a JSON key file, so a Windows service or Task Scheduler job can query BigQuery without an interactive Google sign-in. Interactive OAuth sign-in returns a refresh token you can store and reuse.
Pooling, caching and batches
Python Connector for Google BigQuery pools connections, caches data locally to speed up repeat reads, and sends several update statements to BigQuery as one batch.
Shared API with Devart's other connectors
Python Connector for Google BigQuery uses the same connect, cursor and license calls as Devart's connectors for SQL Server, PostgreSQL, Snowflake and others, so one script can talk to several of them the same way.
Benefits
- Familiar code. The same DB API pattern you use for relational databases.
- Unattended jobs. Service account keys remove the need for a person to sign in.
- One support contract. BigQuery joins your other Devart connectors under one vendor.
Real-World Uses
- Windows reporting server. A scheduled script pulls yesterday's GA4 export tables from BigQuery and saves a summary to SQL Server.
- Moving data between warehouses. A script reads from Snowflake and writes to BigQuery with two Devart connectors and the same code style.
- Analyst notebooks on Windows. A Jupyter user queries BigQuery with plain SQL and DB API parameters.
Who It's For, and Who Should Skip It
The connector suits Windows-based developers who want BigQuery behind a DB API 2.0 interface, especially those already using Devart's other Python connectors.
Skip it on macOS or Linux, in Docker, or on Google Cloud's own Linux services: there's no build for them. Most BigQuery users should start with Google's free google-cloud-bigquery library, which supports every BigQuery feature first, including the Storage Read API and DataFrames.
Usage Guide
- Install with
pip install devart-bigquery-connector on Windows. - After buying, run
devart.bigquery.license.activate("<key>") with the key from the order email or customer portal. - Sign in once with
response = bigquery.signin(), or prepare a service account JSON key. - Connect:
conn = bigquery.connect(ProjectId="my-project", DataSetId="analytics", RefreshToken=response["Refresh Token"]). - Query with
cur.execute("SELECT ... WHERE country = ?", ["US"]) and read rows with fetchall().
Before you buy
BigQuery bills by data scanned, and a DB API connector makes it easy to run broad SELECT * queries from a loop. Keep queries selective and check the cost in Google Cloud. The connector runs only on Windows, and its subscription license needs internet access to activate and check its status.
Where it falls short
- Windows only. No macOS or Linux wheels.
- Free official alternative. Google's client library is open source and cross-platform.
- No asyncio API. It's a synchronous DB API 2.0 driver.
- Per-developer licensing. A team of up to four needs the Team license.
Comparisons
Devart vs. google-cloud-bigquery
Google's library is free, cross-platform and supports every BigQuery feature, including pandas DataFrames and the Storage API. Devart's connector costs $74.95 a year, runs only on Windows, and offers a plain DB API cursor, pooling, local caching and vendor support. Google's library is the default for most people.
Python connector vs. Devart's ODBC driver for BigQuery
Devart's ODBC driver for BigQuery also serves Excel, Power BI and other ODBC tools, and Python through pyodbc. If Python is the only consumer, the Python connector costs less; if reports need BigQuery too, the ODBC driver covers both.
Editions and Pricing
| License | 1-year subscription | Perpetual | Who can use it |
|---|
| Single | $74.95 | $149.95 | One developer at one company site |
| Team | $189.95 | On the order page | Up to 4 developers at one company site |
| Site | $339.95 | On the order page | Unlimited developers at one company site |
Buy Now opens the order for Python Connector for Google BigQuery with a Single license on a 1-year subscription; Team, Site, the perpetual license and 2- or 3-year terms are chosen on the order page. An OEM license for shipping the connector inside your own product is arranged with Devart separately.
FAQs
Does Python Connector for Google BigQuery work on macOS or Linux?
No. Python Connector for Google BigQuery is distributed only as Windows wheels for 32-bit and 64-bit Python, on Windows and Windows Server.
Can Python Connector for Google BigQuery use a service account?
Yes. Python Connector for Google BigQuery 1.4, released October 2026, added service account authentication with a JSON key.
Does it support BigQuery's SQL functions?
Yes. Python Connector for Google BigQuery supports BigQuery's SQL dialects and data types, including aggregate, array, datetime, geography, JSON, navigation, net and hash functions.
Which Python versions does Python Connector for Google BigQuery support?
Python Connector for Google BigQuery 1.4 supports Python 3.10 to 3.15.
Can I ship the connector inside an application I sell?
Only with an OEM license. Python Connector for Google BigQuery's Single, Team and Site licenses cover developing and deploying your own scripts within one company.
Can I get a refund on Python Connector for Google BigQuery?
Possibly. Python Connector for Google BigQuery's order page offers a 30-day money-back guarantee if you're not satisfied, but Devart's ordering FAQ describes refunds within 30 days only for technical problems its support can't fix. Use the 30-day trial to be sure before you buy.
System Requirements
- Python 3.10 to 3.15, 32-bit or 64-bit, with pip
- Windows or Windows Server, 32-bit or 64-bit
- A Google Cloud project with BigQuery and a Google account or service account
- Internet access for activating and checking a subscription license
Conclusion
Devart Python Connector for Google BigQuery fits a Windows team that wants BigQuery behind the same DB API code and Devart license as its other databases, and the new service account sign-in makes it usable for scheduled jobs. For most BigQuery work, especially on Linux or in the cloud, Google's free client library is the better choice. If the Devart approach suits you, try it with the 30-day trial and keep an eye on query costs.
Download Free Trial Buy Now