// the take
The official Google-maintained Python client for BigQuery. This repo is archived — development moved to the google-cloud-python monorepo. If you're starting fresh, you'd install the same `google-cloud-bigquery` package but the source is now elsewhere.
The library itself (not this repo) has solid pandas/Arrow integration via `to_dataframe()` and `to_arrow()`, which makes the BigQuery→local-analysis loop practical. The DB-API 2.0 interface means you can plug it into SQLAlchemy or any tool expecting a standard cursor. OpenTelemetry tracing is baked in, not bolted on. The samples directory is genuinely useful — parameterized queries, partitioned tables, CMEK — real patterns, not toy examples.
This is an archived mirror. Anyone filing issues or PRs here is shouting into the void — the actual project is in a monorepo that's harder to navigate and watch. The client is fundamentally async-unfriendly: `query_job.result()` blocks the thread, and there's no native asyncio support, which matters if you're building anything that queries BigQuery in a web handler. Authentication setup is still a GCP ceremony — ADC works fine locally but breaks in surprising ways in non-GCP cloud environments. Cost visibility is zero — nothing in the client tells you how much a query scanned or what it cost before you run it.
We read the README, the directory tree and a key source file in June 2026. Nobody pays to be listed. How we judge →