On 30 Jun 2014, at 04:56 pm, Nathaniel Smith
A real need, which had also been discussed at length, is a truly performant text IO function (i.e. one using a compiled ASCII number parser, and optimally also a more memory-efficient one), but unfortunately all people interested in implementing this seem to have drifted away (not excluding myself from this)…
It's possible we could steal some code from Pandas for this. IIRC they have C/Cython text parsing routines. (It's also an interesting question whether they've fixed the unicode/binary issues, might be worth checking before rewriting from scratch...)
Good point, last time I was playing with Pandas it was not any faster, but now a 10x speedup speaks for itself. Their C engine does not support generic whitespace separators, but that could probably be addressed in a numpy implementation. Derek