lucene-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Shai Erera (JIRA)" <j...@apache.org>
Subject [jira] Commented: (LUCENE-1614) Add next() and skipTo() variants to DocIdSetIterator that return the current doc, instead of boolean
Date Wed, 27 May 2009 21:23:45 GMT

    [ https://issues.apache.org/jira/browse/LUCENE-1614?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12713748#action_12713748
] 

Shai Erera commented on LUCENE-1614:
------------------------------------

bq. BS2.nextDoc is still needing to check if it's supposed to call initCountingSumScorer?
Can we do this in ctor?

After I made add() private, created a ScorerClauseWrapper (in BQ) and passed to BS2 a list
of SCW (to pass in one call Scorer, req, prohib), I couldn't still call initCountingSumScorer
in the ctor. Then it reminded me that we've had this discussion before - it's related to being
able to ask for topScorer() or not. Reason is, if I call initCount...() in the ctor, it advances
the sub scorers. If the scorer is then used as a topScorer, they may be advanced again, if
BS is used.

So I thought, let's not call nextDoc() in BS's ctor, but that leads to other problems, since
the scorers passed may have called nextDoc() themselves (DisjunctionSumScorer) or may not
(ReqExclScorer).

The decision back then was to leave it as-is, and handle it in LUCENE-1630, when we'll be
able to ask for topScorer. What do you think?

bq. Like what caused it to be added?

Two lines below it, where it will be sent to Collector for collection, since it's < max.
This hits AIOOBE for some Collectors.

bq. I think the last patch would return NO_MORE_DOCS from docID() in ConjunctionScorer if
it determines in ctor that no docs match

Actually, now that you write it I do notice I broke that invariant in CS. If there is even
one scorer that doesn't have any docs, lastDoc is already set to NO_MORE_DOCS and calling
docID before nextDoc will return NO_MORE_DOCS and not -1.

So I think it's a fair relaxation. Just to be clear - this is just a relaxation you're talking
about right? This shouldn't affect any of the existing scorers.

So if we're on the same page, I'll document that relaxation and remove the assert call I've
added to IndexSearcher. But I don't think this should actually change scorers.

> Add next() and skipTo() variants to DocIdSetIterator that return the current doc, instead
of boolean
> ----------------------------------------------------------------------------------------------------
>
>                 Key: LUCENE-1614
>                 URL: https://issues.apache.org/jira/browse/LUCENE-1614
>             Project: Lucene - Java
>          Issue Type: Improvement
>          Components: Search
>            Reporter: Shai Erera
>             Fix For: 2.9
>
>         Attachments: LUCENE-1614.patch, LUCENE-1614.patch, LUCENE-1614.patch, LUCENE-1614.patch,
LUCENE-1614.patch, LUCENE-1614.patch
>
>
> See http://www.nabble.com/Another-possible-optimization---now-in-DocIdSetIterator-p23223319.html
for the full discussion. The basic idea is to add variants to those two methods that return
the current doc they are at, to save successive calls to doc(). If there are no more docs,
return -1. A summary of what was discussed so far:
> # Deprecate those two methods.
> # Add nextDoc() and skipToDoc(int) that return doc, with default impl in DISI (calls
next() and skipTo() respectively, and will be changed to abstract in 3.0).
> #* I actually would like to propose an alternative to the names: advance() and advance(int)
- the first advances by one, the second advances to target.
> # Wherever these are used, do something like '(doc = advance()) >= 0' instead of comparing
to -1 for improved performance.
> I will post a patch shortly

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


Mime
View raw message