3 ms·
I always thought it was interesting that full text search handles misspellings, as opposed to just telling the client that they've misspelled and to clarify the
by endisneigh 4y ago
I always thought it was interesting that full text search handles misspellings, as opposed to just telling the client that they've misspelled and to clarify their intent and retry.
From this it seems like the best solution would be to use SQLite's FTS and replicate the database using something like rqlite, litestream or mvsqlite, and then loadbalance the requests to SQLite.
Given that SQLite is serverless in the classic sense, seems like a nobrainer.
- hardwaresofton 4y agoThis hasn't made the front page of HN (maybe someone should post it), but Litestream actually doesn't do replication that way any more -- I stumbled upon this commit like 2 days ago: https://github.com/benbjohnson/litestream/pull/411 https://github.com/benbjohnson/litestream/pull/411 [EDIT] - I indulged myself: https://news.ycombinator.com/item?id=33204347 https://news.ycombinator.com/item?id=33204347
- avereveard 4y agoPostgres comes with a trigram module. Supports indexing and its much more stable in its result than relying on stemming especially if you're interested in misspellings, the similarity comes with variants with specialized behavior for total similarity, word similarity etc. You might want to give that a chance.