Dear, I wrote a little script (availlable [here](https://gist.github.com/bioinfornatics/2870d9e7dba192114f32ce8299796269)) to compare i/o perf sqlite vs raw tabulated file in python. The result it is a big win for the raw text writer as it is 10 times faster than sqlite writer. While sqlite claims to be [faster than the filesystem](https://www.sqlite.org/fasterthanfs.html). Is it possible to speedup sqlite ? As sqlite does not enforce the [length of a VARCHAR](https://sqlite.org/faq.html#q9) it seem that sqlite can not split rows to insert them by block by guessing the maxmimal size that can take a row. So If I use [SQLITE\_MAX\_LENGTH](https://sqlite.org/limits.html) does the sqlite engine will optimize prepared insert statement with the use of multithread ? Thanks for your insight Have a nice day best regards