Which languages? English only?
Anykind of fuzzy search required?
I don't have experience with big data set but you may to read about GIST/GIN to pick the right index in your case (probably GIST)
But from what I saw when I prepared this post, some people are getting some decent performance with full-text search on dataset like the size wikipedia.
I think that doesn't take long to try Postgres FTS if you are already using it but it may require more investment if you need to move from MySQL to PG.
Thanks for the reply. English only. The only fuzzyness I've told my users about is "phrase here"~20 which would search for the given phrase within 20 words of each other. That helps for the imperfect OCR'd documents. A lot of the stuff we're indexing is best-guess OCR so it's nowhere near perfect. They are boring agreement documents that people have scribbled all over (setting up the redactions) so fuzzy is somewhat interesting to have. But it's not fuzzy in the sense of a typo, it's fuzzy in the sense that it's OCR'd garbage.
I don't have experience with big data set but you may to read about GIST/GIN to pick the right index in your case (probably GIST) But from what I saw when I prepared this post, some people are getting some decent performance with full-text search on dataset like the size wikipedia.
https://wiki.postgresql.org/images/2/25/Full-text_search_in_...
I think that doesn't take long to try Postgres FTS if you are already using it but it may require more investment if you need to move from MySQL to PG.