Qiaochu Liu

@Jackie-Jiang i think there are schemas shared by realtime tables and offline tables. will it be a problem if we drop it and recreate it later?

Will fix the checkstyle, test failures, and add unit tests

> Is there a reason we can't enable this by default and need a config? Hi @KKcorps thanks for reviewing. I think `snapshotEnabled` can be used as a config field...

@Jackie-Jiang pointed that there is a potential issue cause inconsistent query result. #9095 #9101 when reloading the segments, the invalidDocsIndexes can be empty (drop and recreate) for a short period...

@yupeng9 @Jackie-Jiang Do you have any suggestion on when to persist snapshot? I looked into partitionUpsertMetaDataManager and summarized the [scenarios for persisting snapshot](https://docs.google.com/document/d/1guYCag3VoE2148oAj8v1k714qmuhrKI1Z1HNJt2BLqQ/edit#)

We summarized the challenges and thoughts for partial upsert in this [design](https://docs.google.com/document/d/1AhkZPkf4St3hj96IXyNs6NOiItxm3LXw3NAgNbG3oRo/edit#heading=h.z1lc2k9b0z02) Please review cc @Jackie-Jiang @chenboat @yupeng9

ValidDocIds Snapshot management [PR](https://github.com/apache/pinot/pull/9062) and [Pinot doc](https://github.com/pinot-contrib/pinot-docs/pull/127).

After discussion with @Jackie-Jiang @yupeng9 @chenboat We can break down the feature into the following part. - Design doc updates - part 1. When committing segment, update replaceSegment to clean...

The POC was done in #10047 however there are unhandled corner cases. These corner cases was addressed in #10915

can you check if it aligned with segment commit time (replace consuming segment with sealed segment)? it can be addressed by #11964