A Faster Way of Removing Unused Categories in Pandas?

Question

I\'m running some models in Python, with data subset on categories.

For memory usage, and preprocessing, all the categorical variables are stored as category data t

Answer 1

Your problem is in that you are assigning z.get_group(i) to x. x is now a copy of a portion of z. Your code will work fine with this change

for i in z.groups:
    x = z.get_group(i).copy() # will no longer be tied to z
    x.x = x.x.cat.remove_unused_categories()