Closed alexey0308 closed 2 years ago
Hi Alexey,
I finally checked this one out, pandas is indeed much faster, which is a little surprising to me, but hey, cool!
There is only one issue left, I think we need to add header=None
to the pandas reader function, cause otherwise the first line will always be interpreted as a header, which it isn't. I'm just gonna merge and quickly add this myself.
Thanks for finding this performance upgrade!
thanks for the update, it looks like you finished with the application part and finally got time for implementation :)
this is a cherry pick from the csc PR, which is not related to matrix formats