On Mar 27, 2010, at 5:31 AM, MitchK wrote: > > Hello community, > > since I have searched for a solution to get TermPositions in Solr, I became > more aware of the "payload"-features. So I decided to learn more about > payloads. > In the wiki, there is not much said about them, so I will ask here at the > mailing-list. > > It seems like Payloads are some extra-information for tokens, which I can > customize in any way. > For example, I could write a payloadFilter that gives the highest > scoring-factor to the first token and the lowest to the last one. I also > could say "oh, this word is a substantive. Add this as a > payload-information: <substantive>". > > However: How do I use these information at query-time? How can I influence > the scoring in Solr? > I mean, I could write a payload-interpreter (Am I right to do so with > AveragePayloadFunction from Lucene 2.9.1?) for scoring. > So, if I do so, I can switch the scoring of all substantives without > reindexing the payloads by setting there scoring-factor in the schema.xml > (of course this will need some more extra-modifications).
Unfortunately, there is no query time support for this, other than a custom query parser that is posted in JIRA by Erik Hatcher. > > Can anybody tell me more about how to use payloads with Solr? > > For all the others, who want to learn some basic-information about payloads, > I would suggest to read this article from Grant Ingersoll: > http://www.lucidimagination.com/blog/2009/08/05/getting-started-with-payloads/ > > It is a really good tutorial and introduction to this topic. > > Unfortunately, it seems like he has not written anything about how to > integrate this in Solr (I haven't find anything more). Yeah, this is is unfortunate. Would be nice to have both support for payloads and spans in Solr. -Grant
