
Don't mock BigQuery, test on BigQuery: the rewrite of BigTesty
How I test production BigQuery SQL as is, on the real engine, with datasets that live for one test.

Author profile
GDE Cloud | Docker Captain | Head of Data GroupBees | Data | Serverless | IaC | Devops | Java | Python | FP | Clean Code
Browse the latest writing surfaced through DevArt.

How I test production BigQuery SQL as is, on the real engine, with datasets that live for one test.

Serverless DuckDB on Google Cloud, in your own project. Files on GCS, tables in BigQuery (read and write), what I measured, and what I learned.

A real-world ELT pipeline with Dataform and BigQuery — staging and mart layers, JavaScript includes, dynamic tables and views, GitHub sync over SSH and Terraform provisioning.

How I replaced a cloud LLM with a fully local one — same agent, same tools, zero inference cost.

Replace your 200-line bash build scripts with a single declarative Docker Bake file: parallel builds, shared variables, validators, inheritance, matrices, and Compose as Bake input.

Cloud Build runs on ephemeral workers, so your Docker cache vanishes on every run. Here's how to persist it in Artifact Registry with BuildKit's registry cache — uv stops re-downloading dependencies, builds get faster, greener and cheaper.
Advertisement