deepthi912 opened a new pull request, #19683: URL: https://github.com/apache/pinot/pull/19683
## Context `revertSegmentUpsertMetadata()` asks `getPrevKeyToRecordLocationSize()` twice. The first call only compares against zero — "did the revert leave anything behind?" — and returns early in the normal case. The second call, reached only when something was left, needs the actual count for the warning and the inconsistent-rows metric. Counting is free for the on-heap implementations (`ConcurrentHashMap.size()`), so the shape of the API has never mattered. It does for an implementation whose previous-location map is not a map: a store-backed one has to scan its entire keyspace to produce a number, when the question being asked is one a single seek answers. ## Fix Add `isPrevKeyToRecordLocationEmpty()`, defaulting to the existing `getPrevKeyToRecordLocationSize() == 0`, and use it for the first check. Implementations that can answer emptiness more cheaply than a count override it; the on-heap ones keep the default and behave exactly as before. 🤖 Generated with [Claude Code](https://claude.com/claude-code) -- 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]
