These are chat archives for numpy/numpy

Aug 2017
Reese Pathak
Aug 06 2017 19:28
hey, does anyone here have experience working with massive matrices? I'm dealing with two matrices with 30 million rows. The first has 100 columns, the second has 20 columns. They're currently saved as scipy sparse csr_matrices, and saved in the .npz format.
I'm wondering if there is a efficient/any way to take these two matrices and stack them horizontally, so that there are 120 columns. ideally, I'd like them to end up in hdf5 format