Single text index per segment

[xiangfu0]

We recently observed use cases with 100+ columns requiring text indices, which causes the issue of too many open files as current text index implementation is per column.

The proposal here is to allow one shared Lucene Index for Multiple Columns in Pinot.

Changes

• One Lucene index directory per segment, not per column.
• Each text column is represented as a Lucene field within that shared index.
• TextMatch queries can specify which column (i.e., Lucene field) to search.

Challenges

• Meanwhile the down side is also obvious, that whenever there are changes required for adding/removing column index, the process may take longer time for rebuilding the whole index.
• How to handle null/missing value
• shared index dir is a fixed path per segment

#Future extentions

• Advanced queries: Support cross-field Lucene queries if needed later.

[xiangfu0]

cc: @siddharthteotia @Jackie-Jiang @bziobrowski

[bziobrowski]

@xiangfu0 Do you mean something like #15757 ?

[github-actions]

This issue has been automatically marked as stale because it has not had any activity in the last 150 days. It will be closed in 30 days if no further activity occurs. If this issue is still relevant, please leave a comment to keep it open. Thank you for your contributions!