[
https://issues.apache.org/jira/browse/LUCENE-2939?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13002340#comment-13002340
]
Mark Miller commented on LUCENE-2939:
-------------------------------------
P.S. One that is really a bad bug in my mind - we switched this to be the
default and the old Highlighter did not suffer like this in these situations.
Looking back over the email archives, it bit more than a few people. I'm pretty
sure this bug was the impetus of the Fast Vector Highlighter (which is still
valuable if you *really* do want to highlight over every token in your 3
billion word PDF file ;) ).
You pay this huge perf penalty for no gain and no reason. If you are talking
wikipedia size docs, it won't affect you - but for long documents, doing 10
snippets can be prohibitive, with no workaround. That is not a friendly
neighborhood highlighter.
> Highlighter should try and use maxDocCharsToAnalyze in
> WeightedSpanTermExtractor when adding a new field to MemoryIndex as well as
> when using CachingTokenStream
> ----------------------------------------------------------------------------------------------------------------------------------------------------------------
>
> Key: LUCENE-2939
> URL: https://issues.apache.org/jira/browse/LUCENE-2939
> Project: Lucene - Java
> Issue Type: Bug
> Components: contrib/highlighter
> Reporter: Mark Miller
> Assignee: Mark Miller
> Priority: Minor
> Fix For: 3.2, 4.0
>
> Attachments: LUCENE-2939.patch, LUCENE-2939.patch
>
>
> huge documents can be drastically slower than need be because the entire
> field is added to the memory index
> this cost can be greatly reduced in many cases if we try and respect
> maxDocCharsToAnalyze
> things can be improved even further by respecting this setting with
> CachingTokenStream
--
This message is automatically generated by JIRA.
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
