[
https://issues.apache.org/jira/browse/LUCENE-6394?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Robert Muir updated LUCENE-6394:
--------------------------------
Attachment: LUCENE-6394.patch
Updated patch. its a little better, when the 'exclude' part has an
approximation, we will use it.
I need to refactor the logic again though, currently it uses a hack which is
not great for its "advanceExact" logic:
{code}
if (excludeApproximation.advance(candidate.docID()) != candidate.docID()) {
return AcceptStatus.YES;
} else if (!excludeTwoPhase.matches()) {
excludeApproximation.nextDoc(); // move off this doc, we don't match.
return AcceptStatus.YES;
}
{code}
We should just track an int or something instead, moving off the document with
nextDoc() is not as efficient (we might never need it).
> Add two-phase support to SpanNotQuery
> -------------------------------------
>
> Key: LUCENE-6394
> URL: https://issues.apache.org/jira/browse/LUCENE-6394
> Project: Lucene - Core
> Issue Type: Bug
> Reporter: Robert Muir
> Attachments: LUCENE-6394.patch, LUCENE-6394.patch
>
>
> This query is actually a lot like SpanPositionCheckQuery, except it checks
> that each inclusion Spans does not come near the exclusion side.
> Two-phase iteration should just work the inclusion side, deferring positions
> (the overlap checking against exclusion) until necessary.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
