[erlang-questions] Speed of CSV parsing: how to read 1M of lines in 1 second

[Max Lapshin]

On Sun, Mar 25, 2012 at 6:12 PM, Steve Davis
<steven.charles.davis@REDACTED> wrote:
> I know there must be a reason why nobody has suggested:
>
> re:split(Bin, <<",(?=(?:[^\"]*\"[^\"]*\")*(?![^\"]*\"))">>).
>
> ...just not sure what that reason is?
>


Because after you get binaries with encoded floats you need very
expensive translation of binary to float via list.