Skip to content

SNOW-4522850: Make JDBC ingestion documentation self-contained - #4532

Open
sfc-gh-yuwang wants to merge 1 commit into
yuwang/SNOW-4537767-jdbc-type-mappingfrom
yuwang/SNOW-4522850-jdbc-docs
Open

sfc-gh-yuwang wants to merge 1 commit into
yuwang/SNOW-4537767-jdbc-type-mappingfrom
yuwang/SNOW-4522850-jdbc-docs

Conversation

@sfc-gh-yuwang

Copy link
Copy Markdown
Collaborator

Summary

DataFrameReader.jdbc() previously offered placeholder examples and misleading advice about max_workers, schema-inference bypass, and unlimited timeouts. Its docstring now provides a complete setup-to-persistence workflow and explains the behavior implemented by this stack.

  • Add prerequisite SQL for network rules, PASSWORD secrets, external access integrations, role privileges, and uncompressed staged JAR uploads; Notebook/headless session setup; bounded Oracle/PostgreSQL/MySQL/SQL Server examples, Oracle NNE, query/table/options reads, partitions, schema overrides, and persistence.
  • Document current numeric/boolean/timezone/unknown mappings, conservative MySQL BIT handling, current binary limitations, driver buffering/fetch-size caveats, SQL Server subquery ORDER BY, 500-column server limit versus signature length, and blocking-handler timeouts. Link public Snowflake/vendor references and add an entry in the existing RST I/O API page.
  • Add regression checks compiling all six Python example blocks and verifying partial schema overrides preserve inference and source order. No runtime behavior changes; the private-preview marker remains unchanged.

Stack and coordination

Based on exact commit 3e33f580a213954204018a3ce08aeb4cccb9ab63, with PR #4530 (yuwang/SNOW-4537767-jdbc-type-mapping) as the base. This is scoped to SNOW-4522850, not a GA declaration. Query string literals already work in this base, and Oracle timezone values no longer have the stale VARIANT behavior described in the original ticket.

Before integrating the sibling binary/parity work, reconcile the docstring's current binary string-conversion warning (SNOW-4554440), fetch_size=0/schema-inference and other parity semantics (SNOW-4554443). CI and release/preview changes remain with their separate owners. No internal account parameters are advertised as customer controls.

Test plan

  • Datasource unit suite: 86 passed, including both new documentation regressions.
  • Full clean Sphinx HTML build: python3 -m sphinx -E -a -W --keep-going -q -b html -j 4 ...; no Sphinx warnings. Existing Snowpark pandas hybrid-mode Python UserWarning only.
  • Inspect generated JDBC HTML: all six Python blocks render as highlighted code, and setup/usage notes render correctly.
  • Bounded live reads using existing integration fixtures: one-row literal queries passed on Oracle, PostgreSQL, MySQL, and SQL Server.
  • Oracle live checks passed: AES256/SHA256 NNE, session initialization, five-row range partitioning retaining values outside bounds, string/StructType overrides, predicates, DUAL table, format/options, and session-scoped temporary persistence.
  • Commit hooks passed: pyupgrade, isort, black, whitespace/EOF, debug statements, Python AST, license, flake8. git diff --check passed.

Live checks used the existing fixture connection after the preferred local snowpark connection failed with an expired OAuth token. They reused existing source credentials/EAI and uploaded repository JARs to a session stage; no external databases or persistent source tables were provisioned. The temporary persistence check substitutes for creating the documented permanent destination table.

Production TLS certificate/trust configuration for PostgreSQL/MySQL/SQL Server and provisioning/grant SQL were not executed; these are documented against current public vendor/Snowflake references. Live tests used existing fixture connection settings (including SQL Server's test-only certificate bypass). Placeholder examples require deployment-specific hosts, certificates, roles, credentials, and paths.

Generated with Snowflake CoCo

Document source setup, bounded examples, and current conversion and execution constraints so readers can configure ingestion without guessing driver or Snowflake behavior.

Generated with [Snowflake CoCo](https://docs.snowflake.com/en/user-guide/cortex-code/cortex-code)

Co-authored-by: Snowflake CoCo <noreply@snowflake.com>
@sfc-gh-yuwang
sfc-gh-yuwang requested a review from a team as a code owner November 18, 2026 23:51
@sfc-gh-yuwang
sfc-gh-yuwang requested review from sfc-gh-aling, sfc-gh-jzeng and sfc-gh-mayliu and removed request for a team November 18, 2026 23:51
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.

1 participant