Skip to content

docs: AJDA-2839 add BigQuery section to Storage Access docs#986

Open
sykora-ji wants to merge 3 commits into
mainfrom
sykorajiri-AJDA-2839
Open

docs: AJDA-2839 add BigQuery section to Storage Access docs#986
sykora-ji wants to merge 3 commits into
mainfrom
sykorajiri-AJDA-2839

Conversation

@sykora-ji

Copy link
Copy Markdown
Contributor

link to issue

Description

Adds a "Working with the BigQuery Backend" section to the Storage Access (Direct Storage Access) docs page (src/content/docs/data-apps/storage-access/index.md), which previously documented only the Snowflake backend.

Changes:

  • Replaced the top "Snowflake only" caution with a :::note stating Storage Access works on both Snowflake and BigQuery, with a link to the new section.
  • New section documents the two BigQuery-specific differences verified during hands-on testing (parent issue AJDA-2835):
    • Identifier quoting — backticks, quoted per segment (`dataset`.`table`), with the project … has not enabled BigQuery failure mode for the whole-FQN form, plus a Snowflake-vs-BigQuery comparison table.
    • Bucket → dataset name mangling. and - are replaced with _ (e.g. out.c-Test-Data---Customers-Products-Ordersout_c_Test_Data___Customers_Products_Orders); only the dataset name is mangled, the table name keeps its original form.
  • Points users to the Storage table Overview tab (which surfaces the exact Dataset Name and Table Name) as the primary way to find the correct identifiers, with INFORMATION_SCHEMA.SCHEMATA mentioned as a runtime-discovery fallback.
  • Updated the Architecture wording ("not directly with Snowflake" → backend-agnostic) and the Limitations list (removed "Snowflake only", added the SQL-dialect-differs-by-backend note).

Docs-only change. Verified with a clean production build (astro build, 254 pages) and the link/anchor audit; all internal anchors resolve.

Release Notes

  • Justification
    • Direct Storage Access works on BigQuery projects (validated in AJDA-2835), but the docs claimed "Snowflake only", leaving BigQuery users without guidance on the differing SQL dialect — they would hit confusing errors around identifier quoting and dataset names.
    • The docs now explain how to read/write Storage tables from a Data App on a BigQuery project: how to quote identifiers and where to find the correct dataset and table names.
  • Plans for Customer Communication
    • No customer communication needed — documentation-only improvement. The content can optionally feed a changelog note that Direct Storage Access is documented for BigQuery.
  • Impact Analysis
    • No runtime/product impact; only Markdown content under src/content/docs/. No effect on single-tenant setups. Not behind a feature flag.
  • Deployment Plan
    • Continuous deployment — merging to main triggers the GitHub Actions build + aws s3 sync to help.keboola.com.
  • Rollback Plan
    • Fully reversible by reverting the commit and redeploying. Not a one-way-door change.
  • Post-Release Support Plan
    • No monitoring needed. No need to notify the Support team.

@linear

linear Bot commented Jun 24, 2026

Copy link
Copy Markdown

AJDA-2839

@vercel

vercel Bot commented Jun 24, 2026

Copy link
Copy Markdown

The latest updates on your projects. Learn more about Vercel for GitHub.

Project Deployment Actions Updated (UTC)
connection-docs Ready Ready Preview, Comment Jun 25, 2026 10:58am

Request Review

@sykora-ji sykora-ji marked this pull request as ready for review June 24, 2026 13:34
@sykora-ji sykora-ji requested a review from MiroCillik June 24, 2026 13:35

@keboola-pr-reviewer-bot keboola-pr-reviewer-bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Verdict: needs_human (risk 2/5) · profile docs

The PR introduces substantive product-behaviour claims about BigQuery that require human confirmation per policy Section 2.

Concerns:

  • src/content/docs/data-apps/storage-access/index.md: New BigQuery identifier-quoting and dataset-mangling rules are unverified product-behaviour claims.
  • src/content/docs/data-apps/storage-access/index.md: Systematic 'Data App' → 'App' rename not in canonical-term table; needs docs-team confirmation.

Suggested reviewers: @keboola/docs

A two-part `dataset.table` reference works whether quoted per segment
(`` `dataset`.`table` ``) or as a single pair (`` `dataset.table` ``) —
per-segment quoting is not required. The real failure is adding a third
leading segment (the Keboola stage `in`/`out`, or splitting the dotted
bucket ID), which BigQuery resolves as a GCP project ("The project <stage>
has not enabled BigQuery").

Verified live against a real BigQuery project (in_c_shared_bucket.cashier-data):
`ds`.`tbl` and `ds.tbl` both succeed; `out`.`ds`.`tbl` and `out.ds.tbl` fail.

Rewrites the rule, examples, comparison table, and Limitations note; renames
the subsection heading (no internal links referenced the old anchor).
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants