You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@lucene.apache.org by "Erick Erickson (JIRA)" <ji...@apache.org> on 2017/05/19 16:02:04 UTC
[jira] [Comment Edited] (SOLR-10708) Cannot find certain long
values when docValues=true,indexed=false
[ https://issues.apache.org/jira/browse/SOLR-10708?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16017590#comment-16017590 ]
Erick Erickson edited comment on SOLR-10708 at 5/19/17 4:01 PM:
----------------------------------------------------------------
David:
Stupid test. Yes, it fails. With a query parsing error on a negative number. The test is completely invalid so I removed it. Let me dig some more.
But do note that the original problem did not involve negative numbers, so I think there's still an issue here.
was (Author: erickerickson):
David:
Stupid test. Yes, it fails. With a query parsing error on a negative number. The test is completely invalid so I removed it. Let me dig some more.
> Cannot find certain long values when docValues=true,indexed=false
> -----------------------------------------------------------------
>
> Key: SOLR-10708
> URL: https://issues.apache.org/jira/browse/SOLR-10708
> Project: Solr
> Issue Type: Bug
> Security Level: Public(Default Security Level. Issues are Public)
> Affects Versions: 6.7
> Reporter: Erick Erickson
>
> I define a field like this:
> <field indexed="false" name="facet_d" stored="false" type="tlong" docValues="true"/>
> When I index these with a program assigning random longs to the fields searching on facet_d fails to find some values, but _only_ when there are multiple documents. For instance, the value "6850281131226296000" was in this field, but searching on facet_d:6850281131226296000 does not find the document.
> Searching on facet_d:[6850281131226290000 TO 6850281131226297000] works if the range is "wide enough", but I don't have a good sense of how wide that is. Note that the values in that clause are not identical!
> NOTE: this problem does _not_ occur if there's only one document in the corpus. I'll try to make a failing test.
> [~jpountz] suggested that I look at the actual indexed values, but since the document is found in the single-doc case I'll assume that the stored compatibly between the index and query at least.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org