[
https://issues.apache.org/jira/browse/LUCENE-7895?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16074530#comment-16074530
]
Adrien Grand commented on LUCENE-7895:
--------------------------------------
bq. It's all very well saying that these queries are slow, but it still should
be possible to run them if there's no alternative.
Actually I wish we never added {{SpanMultiTermQueryWrapper}}. It creates a
{{SpanTermQuery}} for every matching term without enforcing any limit, which
makes it very trappy: you can easily end up with very slow queries or ever
out-of-memory errors at rewrite time. Can this prefix-inside-phrase problem be
solved at index-time instead with something like edge-ngrams? I'd be ok with
adding hooks to {{QueryBuilder}} that make it easier to handle an index-time
solution of this problem, but if the main use-case is to make it easier to
create a {{SpanMultiTermQueryWrapper}}, then it feels like adding an API only
to enable usage of a trappy query to me.
> Add hooks to QueryBuilder to allow for the construction of MultiTermQueries
> in phrases
> --------------------------------------------------------------------------------------
>
> Key: LUCENE-7895
> URL: https://issues.apache.org/jira/browse/LUCENE-7895
> Project: Lucene - Core
> Issue Type: New Feature
> Reporter: Alan Woodward
> Assignee: Alan Woodward
> Attachments: LUCENE-7895.patch
>
>
> QueryBuilder currently allows subclasses to override simple term query
> construction, which lets you support wildcard querying. However, there is
> currently no easy way to override phrase query construction to support
> wildcards.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
