The feature to load csv is pretty handy if you just need to figure out how to prototype a database from jumbled CSV data without worrying about types and then using SQL instead of messing around too hard on scripts to merge things into a third normal form. From there, export to something like postgres can happen if that's more suited to your use case.
Discussed at the time:<p><i>A Minimalist Guide to SQLite</i> - <a href="https://news.ycombinator.com/item?id=15605669" rel="nofollow">https://news.ycombinator.com/item?id=15605669</a> - Nov 2017 (120 comments)