[
https://issues.apache.org/jira/browse/LUCENE-9450?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17383683#comment-17383683
]
Mayya Sharipova commented on LUCENE-9450:
-----------------------------------------
I am not familiar how taxonomy works, so you adding migration instructions
would be very helpful.
Answering your question:
> If the user has 4 old StoredFields based segments and 1 new BinaryDocValues
> based segment (like the way you just described), does force merging them to 1
> produce a single BinaryDocValues based segment?
We have `FieldInfo` class for each field with a distinct name, and this class
contains information how this field is indexed (for example with
BinaryDocValues). From v 9.0, there is a check that in each segment of the
index each field is indexed the same way, you can't have a field indexed with
BinaryDocValues in one segment, and without BinaryDocValues in another segment.
In 8.x you can still have this option, and when you merge segments to a single
segment, the resulting segment will record this field as indexed with
BinaryDocValues.
This check in v 9.0 is for each for each individual field, so if you transition
4 fields from stored fields to binary doc values, a separate transition needs
to be done for each field.
> reindexing documents as compared to force merging segments to one, so I guess
>the recommended approach for them would be to just reindex.
Reindexing should work as well, if it not an expensive operation for users.
> Taxonomy index should use DocValues not StoredFields
> ----------------------------------------------------
>
> Key: LUCENE-9450
> URL: https://issues.apache.org/jira/browse/LUCENE-9450
> Project: Lucene - Core
> Issue Type: Improvement
> Components: modules/facet
> Affects Versions: 8.5.2
> Reporter: Gautam Worah
> Priority: Minor
> Labels: performance
> Fix For: main (9.0)
>
> Attachments: LUCENE-9450-localrun.py-v1, wip_taxonomy_patch
>
> Time Spent: 3h 50m
> Remaining Estimate: 0h
>
> The taxonomy index that maps binning labels to ordinals was created before
> Lucene added BinaryDocValues.
> I've attached a WIP patch (does not pass tests currently)
> Issue suggested by [~mikemccand]
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org
For additional commands, e-mail: issues-h...@lucene.apache.org
