[
https://issues.apache.org/jira/browse/SOLR-7190?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14711312#comment-14711312
]
Jamie Johnson commented on SOLR-7190:
-------------------------------------
While I understand that DocValuesFacets is more performant at query time I am
wondering if there is an option to support both methods given that the current
implementation pushes the caches deep down into lucene without any real option
for extension. The UnInvertedField allowed me to defer the creation of this
cache until query time and then ultimately control what exactly went into the
cache by specifying a custom cache implementation I was able to create a
separate cache entry for different access levels.
> Remove unused UninvertedField
> -----------------------------
>
> Key: SOLR-7190
> URL: https://issues.apache.org/jira/browse/SOLR-7190
> Project: Solr
> Issue Type: Task
> Reporter: Shalin Shekhar Mangar
> Priority: Minor
> Fix For: 5.2, Trunk
>
>
> I was surprised to find that UninvertedField is no longer used in Solr. The
> only references to UninvertedField is from the fieldValueCache inside
> SolrIndexSearcher and that itself is not used anywhere in SolrIndexSearcher
> except for initialization and regeneration. I can't trace when Solr stopped
> using this class but in any case, we should remove it.
> In a related note, Lucene's DocTermOrds has a copy of the class level
> javadocs of UninvertedField (which extends DocTermOrds). This was done in in
> LUCENE-5666.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
