Fix AGGREGATE expansion in CTEs #12
Conversation
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 7c39936e16
ℹ️ About Codex in GitHub
Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".
| let query = sql.trim().trim_end_matches(';').trim(); | ||
| if let Some(group_by_pos) = find_top_level_keyword(query, "GROUP BY", 0) { | ||
| let start = group_by_pos + "GROUP BY".len(); | ||
| let end = find_first_top_level_keyword( |
There was a problem hiding this comment.
The new find_top_level_keyword will match valid GROUP BY (multiple spaces), but start is calculated as group_by_pos + "GROUP BY".len(), which assumes exactly one space. For inputs like GROUP BY col, start points at the B in BY, so group_by_content becomes BY col and the first extracted column is BY col, which then feeds incorrect group_by_cols into AT correlation and expansion. Consider advancing past the real keyword end (e.g., skip whitespace after GROUP and BY) so arbitrary spacing doesn’t corrupt the parsed columns.
Useful? React with 👍 / 👎.
Summary: expand AGGREGATE inside CTE subqueries before top-level rewrite and add a CTE regression test.
Tests: make test (fails: extension-ci-tools/makefiles/duckdb_extension.Makefile missing).