Author Topic: Transferring large arrays  (Read 471 times)

DougAJ4

  • Newbie
  • *
  • Posts: 27
  • Karma: +0/-0
    • View Profile
Transferring large arrays
« on: March 22, 2014, 11:56:17 PM »
Others here might be interested in my two recent blog posts on transferring data between Excel and Python using Pyxll:

http://newtonexcelbach.wordpress.com/2014/03/16/transfer-of-arrays-tofrom-python-with-pyxll-part1-data-types/
http://newtonexcelbach.wordpress.com/2014/03/22/transfer-of-arrays-tofrom-python-with-pyxll-part-2-speed/

The main messages are:
If you have maths intensive tasks, the Numba compiler can speed things up by factors of up to 500, but only if you use the right data types.
Transferring large arrays from Excel via Pyxll is much faster than reading from Excel in Python.
There are significant differences in the data types depending on whether you use a Pyxll data type, or read an Excel range directly from Python.

Any comments or questions are welcome.