Continuing the discussion from Infer filters - can multiple be applied?:
If there are multiple terms in a query that trigger a nlpFilter, do those filters get and-joined? And if so, is there a way to prevent that / ignore one of the terms? E.g. I’m building an experience for a university where the search “How is financial aid affected by covid-19” is returning no faq results, I think because both “financial aid” and “covid-19” are triggering nlpFilters.
Hi Alex,
You’re correct, multiple NLP filters will be “and” joined, in that only entities that meet all criteria will surface.
Would textSearch work instead for this use case? For fields that have a lot of different values (and are more keyword-like than category-like), that’s typically our recommendation.
Hi Amani,
That’s a great point - originally we had just a few “topics” for FAQs, but that list has grown so much that it really is more like keywords now, so textSearch makes more sense. I suppose the downside is that the facets won’t automatically be checked on the vertical tab, but that’s not a huge drawback.
Thanks!
Alex