Did you add: fq={!geofilt} ??

On 7/3/11 11:14 AM, "Thomas Heigl" <tho...@umschalt.com> wrote:

>Hello,
>
>I just tried up(down?)grading our current Solr 4.0 trunk setup to Solr
>3.3.0
>as result grouping was the only reason for us to stay with the trunk.
>Everything worked like a charm except for one of our queries, where we
>group
>results by the owning user and sort by distance.
>
>A simplified example for my query (that still fails) looks like this:
>
>q=*:*&group=true&group.field=user.uniqueId_s&group.main=true&group.format=
>grouped&sfield=user.location_p&pt=48.20927,16.3728&sort=geodist()
>> asc
>
>
>The exception thrown is:
>
>Caused by: org.apache.solr.common.SolrException: Unweighted use of sort
>> geodist(latlon(user.location_p),48.20927,16.3728)
>> at
>> 
>>org.apache.solr.search.function.ValueSource$1.newComparator(ValueSource.j
>>ava:106)
>> at org.apache.lucene.search.SortField.getComparator(SortField.java:413)
>> at
>> 
>>org.apache.lucene.search.grouping.AbstractFirstPassGroupingCollector.<ini
>>t>(AbstractFirstPassGroupingCollector.java:81)
>> at
>> 
>>org.apache.lucene.search.grouping.TermFirstPassGroupingCollector.<init>(T
>>ermFirstPassGroupingCollector.java:56)
>> at
>> 
>>org.apache.solr.search.Grouping$CommandField.createFirstPassCollector(Gro
>>uping.java:587)
>> at org.apache.solr.search.Grouping.execute(Grouping.java:256)
>> at
>> 
>>org.apache.solr.handler.component.QueryComponent.process(QueryComponent.j
>>ava:237)
>> at
>> 
>>org.apache.solr.handler.component.SearchHandler.handleRequestBody(SearchH
>>andler.java:194)
>> at
>> 
>>org.apache.solr.handler.RequestHandlerBase.handleRequest(RequestHandlerBa
>>se.java:129)
>> at org.apache.solr.core.SolrCore.execute(SolrCore.java:1368)
>> at
>> 
>>org.apache.solr.client.solrj.embedded.EmbeddedSolrServer.request(Embedded
>>SolrServer.java:140)
>> ... 39 more
>
>
>Any ideas how to fix this or work around this error for now? I'd really
>like
>to move from the trunk to the stable 3.3.0 release and this is the only
>problem currently keeping me from doing so.
>
>Cheers,
>
>Thomas


Reply via email to