[
https://issues.apache.org/jira/browse/LUCENE-9207?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17030899#comment-17030899
]
Michael Gibney commented on LUCENE-9207:
----------------------------------------
Thanks, [~romseygeek]. You're right, I was wrong about LUCENE-7398; in its
current state (and any state that doesn't implement backtracking)
SpanNearQuery/SpanOrQuery can potentially miss matches in cases like the one
you describe, even when slop=0. So that's surely still a bug, which would
indeed be addressed by this change (so to be explicit, that makes me +1 to this
change).
Yes, maxBooleanClauses is a good failsafe, but I think it's worth specifically
calling attention to the possibility that for some analyzer configurations and
inputs, this will result in queries failing differently (and more consistently
and transparently) than other queries had failed before (silently missing
matches under certain conditions).
bq.If you have a heavily branching token stream then it's going to produce an
unwieldy query whatever we do, really...
True in some ways; but the characteristics of the implementations do vary
fundamentally, so it's not really 6-of-1, half-dozen-of-another. A complete
nested SpanQuery (as proposed for LUCENE-7398, or analogous Intervals)
implementation has the potential to be significantly more efficient than
MultiPhraseQuery or its analogous Intervals impl (which expand all possible
variants up front).
> Don't build SpanQuery in QueryBuilder
> -------------------------------------
>
> Key: LUCENE-9207
> URL: https://issues.apache.org/jira/browse/LUCENE-9207
> Project: Lucene - Core
> Issue Type: Improvement
> Reporter: Alan Woodward
> Assignee: Alan Woodward
> Priority: Major
> Time Spent: 10m
> Remaining Estimate: 0h
>
> Subtask of LUCENE-9204. QueryBuilder currently has special logic for graph
> phrase queries with no slop, constructing a spanquery that attempts to follow
> all paths using a combination of OR and NEAR queries. Given the known bugs
> in this type of query (LUCENE-7398) and that we would like to move span
> queries out of core in any case, we should remove this logic and just build a
> disjunction of phrase queries, one phrase per path.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
