You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@lucene.apache.org by "Shawn Heisey (JIRA)" <ji...@apache.org> on 2015/09/05 19:15:45 UTC

[jira] [Commented] (SOLR-8012) Field mapping portion of "fl" should be handled in the merger node

    [ https://issues.apache.org/jira/browse/SOLR-8012?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14732050#comment-14732050 ] 

Shawn Heisey commented on SOLR-8012:
------------------------------------

Devil's advocate comment:

What if you have thousands of fields in your index, but you only requested three of them?  Do you want your shards returning all of those fields to the merger, only to have virtually all of that information dropped before being sent to the client?


> Field mapping portion of "fl" should be handled in the merger node
> ------------------------------------------------------------------
>
>                 Key: SOLR-8012
>                 URL: https://issues.apache.org/jira/browse/SOLR-8012
>             Project: Solr
>          Issue Type: Improvement
>          Components: search, SearchComponents - other, SolrCloud
>    Affects Versions: 5.1
>            Reporter: Chirag Gupta
>              Labels: starter
>
> Today "fl" is used for 
> a) filtering the fields and 
> b) mapping the field names to alternate names.
> we should optimize the network traffic by not passing the field mappings to all the nodes and run on the node which is fanning-out and merging the request.
> This impacts the network in the cases where there are 100s of shards to send and the field mapping list is huge



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org