Skip to content

Regenerate tiny TPC-H data with decimals - #366

Draft
bdice wants to merge 1 commit into
rapidsai:mainfrom
bdice:fix/tpch-decimal-test-data
Draft

Regenerate tiny TPC-H data with decimals#366
bdice wants to merge 1 commit into
rapidsai:mainfrom
bdice:fix/tpch-decimal-test-data

Conversation

@bdice

@bdice bdice commented Jul 2, 2026

Copy link
Copy Markdown
Contributor

Summary

  • regenerate the committed SF0.01 TPC-H Parquet fixture using decimal columns
  • keep the fixture aligned with the existing decimal TPC-H table schemas
  • refresh the generated metadata

Root cause

The committed tiny TPC-H files still stored monetary and quantity columns as DOUBLE, while the table schemas declared those columns as DECIMAL(15,2). Registering the decimal schemas over the float fixture caused GPU integration failures from physical/logical type mismatches.

Impact

Local and lightweight integration runs now exercise the same decimal TPC-H types as CI instead of silently using the older float fixture.

Validation

  • verified all nine expected columns are decimal128(15,2) and no fixture columns are floating-point
  • regenerated Presto DDL from the Parquet files and confirmed it matches the committed TPC-H schemas
  • python -m pytest -q benchmark_data_tools/tests/decimal_conversion_test.py
  • ran all 21 Presto GPU TPC-H integration queries locally: 21 passed

@copy-pr-bot

copy-pr-bot Bot commented Jul 2, 2026

Copy link
Copy Markdown

Auto-sync is disabled for draft pull requests in this repository. Workflows must be run manually.

Contributors can view more details about this message here.

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