[ 
https://issues.apache.org/jira/browse/LUCENE-1483?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12655647#action_12655647
 ] 

Michael McCandless commented on LUCENE-1483:
--------------------------------------------

bq. How about those keeping their own reader to filter cache or something? A 
single filter could correctly apply against a single MultiReader before...now 
you would need a different filter for each subreader right?

Such cases would then be caching by sub-reader, right?  That's the benefit of 
this approach.  EG I had been thinking we'd need to fix the recently added 
FieldCacheRangeFilter to also "understand" when it's dealing w/ concatenated 
sub-readers, but in fact with this change that filter is only given the 
sub-readers, one at a time, and so it only asks FieldCache per sub-reader, and 
we automatically get faster reopen performance "for free" (no change to 
FieldCacheRangeFilter required).

Still, I agree this is probably dangerous to suddenly change, since there could 
easily be filters out there that are [illegally] using a docID not 
belonging/corresponding to the reader that was passed in.  So maybe we should 
provide a migration path.  EG, add "allowSubReaders" to Filter, defaulting to 
"return false" so that any external Filter impls still get passed the 
Multi*Reader, and then fix all core/contrib filters to return true from that 
method?

> Change IndexSearcher to use MultiSearcher semantics for sorted searches
> -----------------------------------------------------------------------
>
>                 Key: LUCENE-1483
>                 URL: https://issues.apache.org/jira/browse/LUCENE-1483
>             Project: Lucene - Java
>          Issue Type: Improvement
>    Affects Versions: 2.9
>            Reporter: Mark Miller
>            Priority: Minor
>         Attachments: LUCENE-1483.patch, LUCENE-1483.patch, LUCENE-1483.patch, 
> LUCENE-1483.patch, LUCENE-1483.patch, LUCENE-1483.patch
>
>
> Here is a quick test patch. FieldCache for sorting is done at the individual 
> IndexReader level and reloading the fieldcache on reopen can be much faster 
> as only changed segments need to be reloaded.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to