Remove VectorValues#size() [LUCENE-10176]
- Dominant language
- Java
- Stars
- 3.6k
- Forks
- 1.4k
- Avg merge
- 2d 11h
- Merged PRs (30d)
- 88
Description
This method doesn't seem to be used anywhere except by SimpleTextKnnVectorsReader#search, which uses it in an incorrect way by using it as the total number of hits matching a nearest-neighbor search (it is incorrect because this number might be higher than the number of vectors having a value because of deletes).
---
Migrated from [LUCENE-10176](https://issues.apache.org/jira/browse/LUCENE-10176) by Adrien Grand (@jpountz), updated Feb 16 2022
Contributor guide
Research direction
Start by tracing VectorValues#size() and its use in SimpleTextKnnVectorsReader#search. Check how deletes affect the number of vectors with values, then remove the method and correct the nearest-neighbor search usage; done means the obsolete API is gone and the delete case no longer treats it as the total hit count.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- java
- Domain
- search
- Issue type
- Refactor
- Difficulty
- 3/5
- Estimated time
- 1-2 days
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 38/100