lucene-solr-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Luca Molteni" <voloth...@gmail.com>
Subject Re: How to handle large field values.
Date Wed, 05 Nov 2008 11:51:26 GMT
This worked, thank you very much.

Any idea on how I can help documenting it? Can I write in the wiki?

maybe in

http://wiki.apache.org/solr/SolrConfigXml#head-13e17f74dde0751b8a7cfe539f631d58029b8080

L.M.



2008/11/5 Noble Paul നോബിള്‍ नोब्ळ् <noble.paul@gmail.com>:
> the fl must have the unique id field also.
> because if fl is mentioned it returns only the mentioned one
>
> On Wed, Nov 5, 2008 at 4:36 PM, Luca Molteni <volothamp@gmail.com> wrote:
>> Uhm, this works great when using only one server, because I can
>> specify the fields in the configuration file, but It gives me a nice
>> nullpointer exception when using distributed shards:
>>
>> HTTP Status 500 - null java.lang.NullPointerException at
>> org.apache.solr.handler.component.QueryComponent.returnFields(QueryComponent.java:511)
>> at org.apache.solr.handler.component.QueryComponent.handleResponses(QueryComponent.java:270)
>> at
>>
>>
>>      for (SolrDocument doc : docs) {
>>        Object id = doc.getFieldValue(keyFieldName);
>>        ShardDoc sdoc = rb.resultIds.get(id.toString());
>>        if (returnScores && sdoc.score != null) {
>>          doc.setField("score", sdoc.score);
>>        }
>>        rb._responseDocs.set(sdoc.positionInResponse, doc);
>>      }
>>
>> Any idea?
>>
>> L.M.
>>
>>
>> 2008/11/5 Noble Paul നോബിള്‍ नोब्ळ् <noble.paul@gmail.com>
>>>
>>> the 'fl' parameter can be added to the defaults for your search
>>> handler in solrconfig.xml
>>>
>>> On Wed, Nov 5, 2008 at 3:22 PM, Luca Molteni <volothamp@gmail.com> wrote:
>>> > Hello everybody,
>>> >
>>> > dealing with very large fields, let's say text documents, I found that there
>>> > is a global slowness (on my computer)  in returning those field. Since most
>>> > of the time what we want is an "highlight" value of the field and not the
>>> > entire field, I thought that we can omit these field from the query. I've
>>> > tried two methods:
>>> >
>>> > - Stored = false works very well, the query is faster, but the highlightning
>>> > doesn't work anymore (sigh, I know it's by design).
>>> > - Stored = true and filtering with the "fl" parameters requires me to enter
>>> > all the fields by hand, since the fl query doesn't support a minus operator
>>> > (let's say, all the field withouth my
>>> > veryLargeFieldIDontWantToRetrieveButIWantToSeeAVerySmallPortion).
>>> >
>>> > Strangely, using the "fl" parameter in federeated search with "shards" and
>>> > two different server with different schemas, gave me strange results (no
>>> > results, actually). It all works well using only one shard, but it was very
>>> > difficult to benchmark it.
>>> >
>>> > Any advice? I hope I'm missing something.
>>> >
>>> > Thanks,
>>> >
>>> > L.M.
>>> >
>>>
>>>
>>>
>>> --
>>> --Noble Paul
>>
>
>
>
> --
> --Noble Paul
>
Mime
View raw message