Hi,

I would first of all try the most simple solution, namely re-index bug 
documents on every change, even if they occur frequently. Solr can handle quite 
some update load, so I would not call it impossible before I've tried.

If all you need the external value for is boosting score, then you could use 
ExternalFileField which is a .txt file which you can update when you want 
without reindexing. But these cannot be queried or returned in results, only 
used for boosting and sorting.

The next option is Join. Solr may perform very simple join on a "foreign key", 
in which case you may update the frequently-changing fields in their own 
separate small documents, and join them in at query time. There are limitations 
to this approach though.

Finally, in 4.0 we get DocValues/Column Stride Fields, which is meant for 
"external" values which change frequently. The idea is that these will be 
updateable without reindexing, but I don't think they are yet. See 
http://www.slideshare.net/lucenerevolution/willnauer-simon-doc-values-column-stride-fields-in-lucene
 for more.

--
Jan Høydahl, search solution architect
Cominvent AS - www.cominvent.com
Solr Training - www.solrtraining.com

On 11. okt. 2011, at 21:47, Ikhsvaku S wrote:

> Help pls sirs :)
> 
> On Tue, Oct 11, 2011 at 1:29 PM, Ikhsvaku S <ikhsv...@gmail.com> wrote:
> 
>> Hi,
>> 
>> We were recently investigating use of Solr for querying & indexing our bug
>> database. We are very happy and most of the fields could be indexed
>> straightforward. But there are some of the fields that cant be indexed as
>> they are changed all the time and we want to incorporate that too in solr
>> query. Even if solr doesnt index that external field it instead accepts the
>> query and calls the external script/db/webservice to get value of the field
>> and proceed with rest...
>> 
>> Like it can allow queries with these external fields but instead of
>> checking it in its own indexes - queries external database/web-service and
>> then return results...?
>> 
>> Eg. userid_s:a* AND my_external_field:a* AND title_s:a*
>> 
>> 
>> Here my_external_field is fetched from webservice/rdbms calls and returns
>> the set of unique doc ids
>> 
>> 
>> Thanks
>> 

Reply via email to