andygrove opened a new pull request, #5117: URL: https://github.com/apache/datafusion-comet/pull/5117
## Which issue does this PR close? Cherry-pick of #5085 (merged to `main` as `ce84a24c1`) onto `branch-1.0`. Documentation-only; each of the referenced `correctness`+`bug` issues remains open until fixed. ## Rationale for this change One of the requirements for the 1.0.0 release is that we document all known correctness issues. #5085 landed that documentation on `main` after `branch-1.0` was cut, so the 1.0.0 release branch is missing it. Users on 1.0.0 can hit these divergences silently unless the compatibility guide published from `branch-1.0` describes them. ## What changes are included in this PR? A straight `git cherry-pick -x` of `ce84a24c1`. The diff is byte-identical to the commit on `main`; there were no conflicts and no adaptation was needed. Expression-level notes surfaced via `getCompatibleNotes` / `getIncompatibleReasons` on the auto-generated per-category compat pages: | Expression | Issue | Nature | |-----------|-------|--------| | `Round` | #5070 | `LongType` returns `0` at `scale <= -19` instead of overflowing | | `IntegralDivide` | #5065 | `Long.MinValue div -1` silently returns wrapped value under ANSI | | `MakeDecimal` | #5066 | Native path drops `fail_on_error`, so ANSI overflow returns `NULL` | | `NextDay` | #5073 | ANSI error surfaces as `CometNativeException` not `SparkIllegalArgumentException` | | `MakeDate` | #5073 | ANSI error surfaces as `CometNativeException` not `SparkDateTimeException` | | `MapFromArrays`, `MapFromEntries` | #4680 | Per-element `NULL` key not rejected (Spark throws) | | `GetJsonObject` | #4947 | Duplicate keys: native returns last, Spark returns first | Static sections in `compatibility/index.md`: - ANSI-mode error classes and messages (#5071, #5072, #5073, #4967) - Known result-value divergences: boolean-to-decimal cast eval-mode (#5068), `CAST(string AS boolean)` ISO controls (#4959), `RANGE` window boundary overflow (#5022), `null IN ()` legacy behavior (#4786) - Object store cache key drops the ABFS container (#4993) `compatibility/scans.md`: expands the datetime-rebase note (#5010) to cover any file written with `datetimeRebaseModeInWrite=LEGACY`, note that filters and aggregates also disagree, that `spark.sql.parquet.datetimeRebaseModeInRead` and the file-level `org.apache.spark.legacyDateTime` metadata are ignored, and that `spark.comet.exceptionOnDatetimeRebase` is currently dead code. ## How are these changes tested? Not applicable: this PR only touches documentation strings and static markdown, and is an unmodified cherry-pick of a commit already reviewed and merged to `main`. The per-category compat pages are regenerated in CI from the `getCompatibleNotes` / `getIncompatibleReasons` values. -- 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]
