Pandas: Is there a way to use something like 'droplevel' and in process, rename the other level using the dropped level labels as prefix/suffix?
Use list comprehension
for set new column names:
df.columns = df.columns.map('_'.join)
Or:
df.columns = ['_'.join(col) for col in df.columns]
Sample:
df = pd.DataFrame({'A':[1,2,2,1],
'B':[4,5,6,4],
'C':[7,8,9,1],
'D':[1,3,5,9]})
print (df)
A B C D
0 1 4 7 1
1 2 5 8 3
2 2 6 9 5
3 1 4 1 9
df = df.groupby('A').agg([max, min])
df.columns = df.columns.map('_'.join)
print (df)
B_max B_min C_max C_min D_max D_min
A
1 4 4 7 1 9 1
2 6 5 9 8 5 3
print (['_'.join(col) for col in df.columns])
['B_max', 'B_min', 'C_max', 'C_min', 'D_max', 'D_min']
df.columns = ['_'.join(col) for col in df.columns]
print (df)
B_max B_min C_max C_min D_max D_min
A
1 4 4 7 1 9 1
2 6 5 9 8 5 3
If need prefix
simple swap items of tuples:
df.columns = ['_'.join((col[1], col[0])) for col in df.columns]
print (df)
max_B min_B max_C min_C max_D min_D
A
1 4 4 7 1 9 1
2 6 5 9 8 5 3
Another solution:
df.columns = ['{}_{}'.format(i[1], i[0]) for i in df.columns]
print (df)
max_B min_B max_C min_C max_D min_D
A
1 4 4 7 1 9 1
2 6 5 9 8 5 3
If len
of columns is big (10^6), then rather use to_series
and str.join
:
df.columns = df.columns.to_series().str.join('_')
Using @jezrael's setup
df = pd.DataFrame({'A':[1,2,2,1],
'B':[4,5,6,4],
'C':[7,8,9,1],
'D':[1,3,5,9]})
df = df.groupby('A').agg([max, min])
Assign new columns with
from itertools import starmap
def flat(midx, sep=''):
fstr = sep.join(['{}'] * midx.nlevels)
return pd.Index(starmap(fstr.format, midx))
df.columns = flat(df.columns, '_')
df