Hello Quynh, Solr has support for external file fields [1]. They are a simple key=float based text file where key is ID, and the float can be used for boosting/scoring documents. This is a much simpler approach than using a separate collection. These files can be reloaded every commit and are really easy to use. We use them for boosting documents by their popularity.
Hope that helps, Markus [1] https://lucene.apache.org/solr/guide/6_6/working-with-external-files-and-processes.html -----Original message----- > From:Huynh, Quynh <quynh.t.hu...@nordstrom.com> > Sent: Thursday 5th April 2018 22:50 > To: solr-user@lucene.apache.org > Cc: Collazo, Carlos <carlos.coll...@nordstrom.com>; Ganesan, VinothKumar > <vinothkumar.gane...@nordstrom.com> > Subject: Storing Ranking Scores And Documents In Separate Indices > > Hey Solr Community, > > We have a collection of product documents that we’d like to add fields to > with ranking scores generated by our data scientists. > > Two options we’re considering is to either: > - Have a separate index that contains all the documents from our > product index, but with these additional ranking fields > - Have an index with just the score fields and a numerical key to > represent the product that would require a separate lookup > > We wanted to know if any Solr users with a similar problem has tried either > of those options (and the performance implications you faced), or had a > different approach to structuring documents in separate collections, where > the only difference between the documents was the ranking fields. > > > Thanks! > Quynh >
