[
https://issues.apache.org/jira/browse/LUCENE-1614?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12710855#action_12710855
]
Michael McCandless commented on LUCENE-1614:
--------------------------------------------
{quote}
> I wonder if instead of returning -1 when the iteration is done, we should
> return Integer.MAX_VALUE?
The idea was to return a negative value and then compare the returned value to
>= 0 for better performance. If we return MAX_VAL we'll need to compare to
MAX_VAL, which is less efficient, CPU wise.
{quote}
If you always check the returned result, right.
But with BooleanScorer/2, and likely any scorer that invokes multiple
sub-scorers, switching to Integer.MAX_INT as the sentinel would allow
us to *not* test every sub-scorer's returned result from
nextDoc/advance. Ie, because the docID moved forward (vs -1, which
moved backwards), it's a "natural" fit for the main scorer's normal docID
processing. So, far fewer if's (one per subscorer) are in the
innermost loop.
Marvin, what's your plan for Lucy's sentinel value for DISI/Scorer?
> 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
>
>
> 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-unsubscr...@lucene.apache.org
For additional commands, e-mail: java-dev-h...@lucene.apache.org
