[
https://issues.apache.org/jira/browse/LUCENE-5069?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13688405#comment-13688405
]
Robert Muir commented on LUCENE-5069:
-------------------------------------
{quote}
With this info in FieldInfo we could automatically select the right precision
step for each atomic reader processed while the query runs.
{quote}
The problem is its too late: QueryParser/Query are independent of readers: so
they dont know to generate the correct query (e.g. NumericRangeQuery instead of
TermRangeQuery) in the first place!
So this issue misses the forest for the trees, sorry, -1 to a halfass schema
that brings all of the problems of a schema and none of the benefits!
> Can/should we store NumericField's precisionStep in the index?
> --------------------------------------------------------------
>
> Key: LUCENE-5069
> URL: https://issues.apache.org/jira/browse/LUCENE-5069
> Project: Lucene - Core
> Issue Type: Improvement
> Reporter: Michael McCandless
>
> I was just helping a user (buzzkills) on IRC on why NumericRangeQuery was
> failing to hit the expected docs ... and it was because s/he had indexed with
> precStep=4 but searched with precStep=1.
> Then we wondered if it'd be possible to somehow catch this, e.g. we could
> maybe store precStep in FieldInfo, and then fail at search time if you use a
> "non-matching" precStep?
> I think you can index fine and then search on a multiple of that? E.g., I
> can index with precStep=2 but search with precStep=8? But indexing with
> precStep=4 and searching precStep=1 won't work ...
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
