How about perf if you dynamically create 5000 fields ? Bill Bell Sent from mobile
> On Sep 14, 2014, at 10:06 AM, Erick Erickson <erickerick...@gmail.com> wrote: > > Dynamic fields, once they are actually _in_ a document, aren't any > different than statically defined fields. Literally, there's no place > in the search code that I know of that _ever_ has to check > whether a field was dynamically or statically defined. > > AFAIK, the only additional cost would be figuring out which pattern > matched at index time, which is such a tiny portion of the cost of > indexing that I doubt you could measure it. > > Best, > Erick > > On Sun, Sep 14, 2014 at 7:58 AM, Saumitra Srivastav > <saumitra.srivast...@gmail.com> wrote: >> I have a collection with 200 fields and >300M docs running in cloud mode. >> Each doc have around 20 fields. I now have a use case where I need to >> replace these explicit fields with 6 dynamic fields. Each of these 200 >> fields will match one of the 6 dynamic field. >> >> I am evaluating performance implications of switching to dynamicFields. I >> have tested with a smaller dataset(5M docs) but didn't noticed any indexing >> or query performance degradation. >> >> Query on dynamic fields will either be faceting, range query or full text >> search. >> >> Are there any known performance issues with using dynamicFields instead of >> explicit ones? >> >> >> >> >> -- >> View this message in context: >> http://lucene.472066.n3.nabble.com/Solr-Dynamic-Field-Performance-tp4158737.html >> Sent from the Solr - User mailing list archive at Nabble.com.
