Question

I’m reading a CSV, and I’d like to make one of the columns a Categorical, with my own ordering. How do I do that? The three labels are ‘read’, ‘write’ and ‘mixed’. Here are some things that don’t work:

Categorical(my_csv.rw, ['read', 'write', 'mixed'])
    ValueError: invalid literal for long() with base 10: 'mixed'

Categorical(my_csv.rw, Index(['read', 'write', 'mixed']))
    ValueError: invalid literal for long() with base 10: 'mixed'

Categorical(['read', 'mixed', 'write'], Index(['read', 'write', 'mixed']))
    ValueError: invalid literal for long() with base 10: 'mixed'

Categorical.from_array(['read', 'mixed', 'write']) # Levels in wrong order

So, how should I do it?

Answer 1

I believe the levels are ordered by the labels:

In [38]: a.levels
Out[38]: Index([mixed, read, write], dtype=object)

In [39]: a.labels
Out[39]: array([1, 0, 2])