kokhlo commented on PR #43682: URL: https://github.com/apache/superset/pull/43682#issuecomment-5468028633
Verified the aggregate/histogram paths and the updated boxplot expectations — consistent with the fixture (the NULL region row now forms its own group, 4→5). One follow-up worth noting for the same post-processing pipeline: pivot (superset/utils/pandas_postprocessing/pivot.py:325) calls df.pivot_table(...) with dropna=drop_missing_columns, which only governs all-NaN columns. NULL keys in the index are still dropped by pivot_table, so a query whose post-processing is aggregate → pivot will now keep the NULL group in the aggregate output but silently lose the same rows again at the pivot step. Charts that pivot (most table/pivot-table visuals) will not see the fix; only the direct aggregate/histogram/boxplot paths do. Not a blocker for this PR — the dropna=False fix is correct where it is applied — but the sibling path probably wants the same treatment (pivot_table has no direct index-dropna flag; the usual approach is filling NULL index keys before the pivot, the way the codebase already handles NULL preservation structurally in _restore_dropped_metric_columns). Happy to take that follow-up separately if maintainers want it. -- 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]
