Hi Ufuk, Rahul, thanks for your support. My collection is quite small (100k docs) so adding uninvertible=true and reindexing all the docs did the trick. With such a small index the performance is still good.
Kind regards, Vincenzo On Mon, Jun 16, 2025 at 6:32 PM ufuk yılmaz <uyil...@vivaldi.net.invalid> wrote: > > If I understand correctly after the values are analyized/tokenized by > the solr.TextField and then copied into the string/strings field Am I > right? > > Not really, copyField directive copies the incoming raw data during > indexing to a second field which can have a completely different type, > analyzer and tokenizer, text processing happens afterwards. If the targeted > field is something that supports docValues, such as a string, faceting and > sorting on that field is much more efficient. Usually what you do is to > search on the original field but facet/sort on the second field, to repeat, > this is only worth to implement if you care about the performance benefit > of docValues. To solve your original problem just adding uninvertible=true > and reindexing would be enough I imagine. > > -Ufuk > > — > > > On Jun 16, 2025, at 17:51, Vincenzo D'Amore <v.dam...@gmail.com> wrote: > > > > If I understand correctly after the values are analyized/tokenized by > > the solr.TextField and then copied into the string/strings field Am I > > right? > > -- Vincenzo D'Amore
