Hoss Man commented on SOLR-11942:

Another thought: in general it might be work adding a "joining and grouping" 
type document, giving a high level overview of the concepts/usecases people 
might have and how to solve them, where we note the general caveats of why 
"distributed grouping/joining" is a hard problem and how things like the 
paralllel SQL APIs can sometimes be of use in these sitautions


> need better ref guide docs on join (and grouping) limitations in solr cloud
> ---------------------------------------------------------------------------
>                 Key: SOLR-11942
>                 URL: https://issues.apache.org/jira/browse/SOLR-11942
>             Project: Solr
>          Issue Type: Improvement
>      Security Level: Public(Default Security Level. Issues are Public) 
>          Components: documentation
>            Reporter: Hoss Man
>            Priority: Major
> on irc, a user asked about joining in solr cloud and i realized that the 
> "JoinQParser" docs don't really clarify what does/doesn't work in cloud.  
> There is a section on "Joining Across Collections" which points out some 
> limitations in cloud mode, and the "Result Grouping" page has a brief note 
> that "a...ll documents in each group must be co-located on the same shard in 
> order for accurate counts to be returned. Document routing via composite keys 
> can be a useful solution in many situations."  But there is no similar 
> warning for using the join parser n a single collection.
> We should probably beef up the warning on the "Result Grouping" section to 
> give an example of what does/doesn't work, and cross link to that from the 
> "join" parser docs? (ie: "Note that when using multiple shards, the {{join}} 
> parser has similar limitations as  {{the GroupComponent regarding co-located 
> documents|LINK_TO_OTHER_DOC}} )

This message was sent by Atlassian JIRA

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

Reply via email to