[
https://issues.apache.org/jira/browse/SOLR-9696?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Dawid Weiss resolved SOLR-9696.
-------------------------------
Resolution: Won't Fix
This seems to be implemented already as we only query the fields we cluster on.
{code}
SolrDocumentList solrDocList =
SolrPluginUtils.docListToSolrDocumentList(
results.docList, rb.req.getSearcher(),
engine.getFieldsToLoad(rb.req), docIds);
Object clusters = engine.cluster(rb.getQuery(), solrDocList, docIds,
rb.req);
{code}
> Keep references to only those fields required for clustering, not entire
> documents
> ----------------------------------------------------------------------------------
>
> Key: SOLR-9696
> URL: https://issues.apache.org/jira/browse/SOLR-9696
> Project: Solr
> Issue Type: Improvement
> Security Level: Public(Default Security Level. Issues are Public)
> Reporter: Dawid Weiss
> Assignee: Dawid Weiss
> Priority: Minor
>
> As reported by user:
> "We face some strange performance issues with respect to memory while
> clustering calls are initiated. We are clustering around 1000 documents and
> these documents are fairly big documents (700 MB in size in unzipped state)
> When we were exploring, we found that the memory consumption came down when
> we add a field to the fl parameter of solr (the field is a double field)."
> I don't know why adding a double field to fl would impact memory consumption,
> but it'd be good to verify we don't hold to too much stuff inside the
> clustering plugin (if not needed).
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: dev-h...@lucene.apache.org
