dwsmith1983 commented on PR #5365: URL: https://github.com/apache/datafusion-comet/pull/5365#issuecomment-5519861917
Thanks for the depth here, and especially for independently verifying the sigma port over 212k strings before asking for the split. Agreed on the restructure, and #5602 landing made most of it the obviously right call. Here is where things stand: 1. The case folding stack is gone entirely. JvmCaseTables, the schema_adapter port, the three proto fields, the parity suite, and the QueryContextInternerSuite pin are all dropped in favor of main's name_fold implementation, which the Delta scan inherits automatically through the shared parquet path. That also dissolves the #5495 concern, since #5602 owns that fix now. 2. The three small fixes moved to #5653 (object store variable expansion, constant metadata uniquification, dead JNI removal). The first two are prerequisites of the contrib, so this branch is based on that PR and shows its commits until it merges. 3. The duplicate field and field id semantics are going out as their own PR shortly, re-derived on top of #5602's folding. Main's version already covers the id-routed duplicate exemption and the case insensitive struct duplicate error, so that PR carries only the shield generalization, the duplicate requested field id error, and a last-field-wins fix for case sensitive duplicate names that main currently resolves first-wins. What remains here is the Delta contrib plus the refactors it genuinely needs, per your list. The branch is rebased on current main, both suite profiles pass, and QueryContextInternerSuite passes on the default configuration again. The description is updated to match. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
