piRSquared piRSquared - 1 month ago 23
Python Question

access pandas axes by axis number

consider the

pd.Panel
pn
and
pd.DataFrame
df


import pandas as pd
import numpy as np

pn = pd.Panel(np.arange(27).reshape(3, 3, 3), list('XYZ'), list('abc'), list('ABC'))

pn.to_frame().rename_axis('item', 1).unstack()


enter image description here

df = pd.DataFrame(np.arange(9).reshape(3, 3), list('abc'), list('ABC'))

df


enter image description here




I can access the
items
axis of
pn
with
pn.items
and the
columns
axis of
df
with
df.columns
.

Question

But how do I get the
items
axis of
pn
with the number
0
and
minor_axis
axis with the number
2
? With as many methods that accept the parameter
axis=0
, I'm imagining there is a straight forward way to access axes via number.

What I've done

Custom function

def get_axis(obj, axis=0):
if isinstance(obj, pd.Panel):
m = pd.Series(['items', 'major_axis', 'minor_axis'])
else:
m = pd.Series(['index', 'columns'])
return obj.__getattribute__(m.loc[axis])

print(get_axis(pn, 2))
print(get_axis(pn, 1))
print(get_axis(pn, 0))
print(get_axis(df, 1))
print(get_axis(df, 0))





Index([u'A', u'B', u'C'], dtype='object')
Index([u'a', u'b', u'c'], dtype='object')
Index([u'X', u'Y', u'Z'], dtype='object')
Index([u'A', u'B', u'C'], dtype='object')
Index([u'a', u'b', u'c'], dtype='object')

Answer

Use .axes to return a list of the index axis labels as well as the column axis labels and then you could access it via slice-notation.

pn.axes
#[Index(['X', 'Y', 'Z'], dtype='object'),
# Index(['a', 'b', 'c'], dtype='object'),
# Index(['A', 'B', 'C'], dtype='object')]

Then, you could provide the slices to retrieve the objects:

pn.axes[0]
#Index(['X', 'Y', 'Z'], dtype='object')

df.axes[0]
#Index(['a', 'b', 'c'], dtype='object')