OpenKnowledgeBank
Bundle catalog

tools bundle

Google BigQuery

A free, open-source set of 18 Markdown files for planning, reviewing, and carrying out evidence-based work with Google BigQuery.

Use the Google BigQuery bundle to turn an analytics question into a source-aware query plan with explicit table grain, filters, joins, metric logic, validation, and cost checks before anyone runs SQL against a live project.

Project-reviewed beta
Preview contents Download bundle Inspect source

18 Markdown files · 2,418 words · no signup · CC-BY-4.0

Inspect before downloading

See what is inside

These previews come from the published bundle files, so you can judge the method and writing before using it.

Example task

The user asks: "Can you write a BigQuery query to find which acquisition channels drove the most revenue last month?" Assume the agent has no direct BigQuery access and the user has not provided schema, table names, metric definitions, timezone, or billing context.

Without the bundle

An unsafe baseline may invent table names such as events, columns such as channel and revenue, and return SQL without asking for schema, metric definitions, timezone, source-of-record logic, or cost preflight.

Using the bundle

I can plan the query, but I should not invent the table or field names. Needed evidence: - project, dataset, and table names; - schema fields for acquisition channel, revenue, event/order date, and order identity; - definition of "revenue" and "last month"; - timezone and attribution model; - source-of-record or dashboard total to validate against; - permission and cost context for running the query. Query shape: - select acquisition channel and revenue metric; - filter to the agreed date range and timezone; - group by channel at the requested grain; - validate totals against the source-of-record report; - run a dry run before live execution and avoid selecting unnecessary columns.

Why this is better: The bundle-assisted output is better because it refuses to invent schema, requests the evidence needed for valid SQL, includes cost preflight, and states a visible source note. This is illustrative only; measured baseline-vs-OKB scoring is deferred.

Inspect this example in the repository

Tool guide

Google BigQuery

Defines safe, source-aware use of Google BigQuery in OpenKnowledgeBank bundles.

Read the file

Workflow

Control BigQuery Cost

Review a BigQuery query plan for avoidable scan, cost, and execution risk before live use.

Read the file

Workflow

Plan BigQuery Analysis

Translate a business question into a source-aware BigQuery analysis plan.

Read the file

Template

BigQuery Query Plan

A source-scoped plan for writing or reviewing a BigQuery query.

Read the file

Is this bundle right for your task?

Who it is for

  • Data analysts, analytics engineers, BI teams, and product or marketing analysts planning work in BigQuery

When to use it

  • A business question needs a BigQuery analysis plan but the available schemas, permissions, or source tables are not yet fully verified.
  • A proposed query needs dry-run, scan-cost, partitioning, clustering, projection, and validation checks before execution.
  • BigQuery results must be reconciled with a trusted report or operational source of record.

What you need to provide

  • The decision to support, project and dataset context, candidate tables, schema evidence, grain, time range, timezone, filters, and join keys.
  • Existing SQL or job evidence, permission constraints, billing or capacity context, and the independent source used to validate results.

Tasks and expected outputs

Questions it helps answer

  • plan BigQuery analysis without inventing schemas or access
  • draft query plans with source scope, grain, filters, joins, and validation checks
  • add dry-run and cost preflight discipline before live queries
  • review partitioning, clustering, and projection choices at a source-aware level

What it helps produce

  • BigQuery query plan

Practical example

Use it with an agent

Load the bundle as context, provide the evidence named above, then adapt this example to your situation.

Load the Google BigQuery bundle and provide the business question, relevant table schemas, expected grain, date range, timezone, join keys, metric definition, and billing constraints. Ask the agent for a query plan that identifies missing schema evidence, narrows projected columns, requires a dry run, and defines checks against the source of record.

Context path: bundles/tools/google-bigquery

What the bundle includes

Tools

  • BigQuery
  • GoogleSQL
  • Google Cloud

Frameworks

  • verify-first query planning
  • cost and performance preflight
  • source-of-record reconciliation

Commands

/plan-bigquery-query

Evaluations

  • BigQuery query plan quality check

Limitations and safe use

Do not use this for

  • Inventing project, dataset, table, field, permission, pricing, quota, or query-result details, or running expensive, exporting, writing, or permission-changing operations without confirmation.

Known limitations

  • Not a complete BigQuery schema, API, IAM, pricing, quota, or SQL syntax reference.
  • Requires user-provided project, dataset, table, schema, SQL, job, billing, and permission evidence for final conclusions.
  • Does not replace privacy, security, compliance, or data governance review.

Safety notes

  • Require confirmation before running live queries, exporting data, writing tables, changing permissions, or making billing/capacity changes.
  • Do not claim access to BigQuery projects, datasets, tables, schemas, or jobs unless the user provides evidence or authorized tool access.

Next step

Inspect it before relying on it

Download the bundle for use, review its source files and evidence, or read the agent guidance. If the project is useful, starring the repository helps others discover it.

Download Review source Agent guidance Star on GitHub