2 ms·
This is something that grows slowly on you, and it is a very hairy problem. By the time you realize the scope of your problem, finding a library that can be sho
by Someone 10y ago
This is something that grows slowly on you, and it is a very hairy problem. By the time you realize the scope of your problem, finding a library that can be shoe-horned into your code and doesn't have huge regressions on your data is hard.
Also, assuming a robust generic solution exists, it will almost certainly be slower than a customized one. It is very tempting to think "but our data is relatively clean, we don't need the full feature set of that library that makes imports take an hour longer to run".