Hi Rick, I don't think the issue is BM25 vs TFIDF (the old similarity), it seems more due to the "matching" logic.
you are asking to match: "(Action AND Technical AND Temporaries AND t/a AND CTR AND Corporation)" This (in theory) means that you want to retrieve **only** the documents that contain **all* the terms (so just the first) - so bm25 seems to do the right thing. Solr allows you to "relax" the requirement, for example using the "mm" parameter (see [1] ), could you try to use mm with bm25 and see if it solves your problem? Are you sure that when you use tfidf you are only changing the similarity and not something else? Can you try to enable the debug mode (&debugQuery=true) and check in the response how the query is processed? (feel free to post here) [1] https://lucene.apache.org/solr/guide/6_6/the-dismax-query-parser.html#TheDisMaxQueryParser-Themm_MinimumShouldMatch_Parameter From: solr-user@lucene.apache.org At: 02/26/18 22:39:43To: solr-user@lucene.apache.org Subject: SOLR Similarity Difference I’m converting SOLR 4.10.2 to SOLR 7.1 I have the following three strings in both SOLR cores Action Technical Temporaries t/a CTR Corporation Action Technical Temporaries Action Technical Temporar If I search IDX_CompanyName: (Action AND Technical AND Temporaries AND t/a AND CTR AND Corporation) Under 4.10.2 I see all three in the results Under 7.1, with the default BF25 similarity. I only see the first result Someone on the list suggested that make 7.1 to go back to the similarity factory used in 4.10.2 that I add the following to the schema.xml. <similarity class="org.apache.solr.search.similarities.ClassicSimilarityFactory"></similarity> That brings all three results. But my boss would prefer that we don’t use the older similarity factory. Is there some setting other than similarity factory that will make 7.1 include these documents without changing the query? Thanks, Rick Hodder Information Technology Navigators Management Company, Inc. 83 Wooster Heights Road, 2nd Floor Danbury, CT 06810 (475) 329-6251