Pandas 'partial melt' or 'group melt'

Question

I have a DataFrame like this

>>> df = pd.DataFrame([[1,1,2,3,4,5,6],[2,7,8,9,10,11,12]], 
                      columns=['id', 'ax','ay','az','bx','by','bz'])
>>> df
   id  ax  ay  az  bx  by  bz
0   1   1   2   3   4   5   6
1   2   7   8   9  10  11  12

and I want to transform it into something like this

   id name   x   y   z
0   1    a   1   2   3
1   2    a   7   8   9
2   1    b   4   5   6
3   2    b  10  11  12

This is an unpivot / melt problem, but I don't know of any way to melt by keeping these groups intact. I know I can create projections across the original dataframe and then concat those but I'm wondering if I'm missing some common melt tricks from my toolbelt.

Answer 1

Not melt wide_to_long with stack and unstack

pd.wide_to_long(df,['a','b'],i='id',j='drop',suffix='\w+').stack().unstack(1)
Out[476]: 
drop   x   y   z
id              
1  a   1   2   3
   b   4   5   6
2  a   7   8   9
   b  10  11  12

Answer 2

Set_index, convert columns to multi index and stack,

df = df.set_index('id')
df.columns = [df.columns.str[1], df.columns.str[0]]
new_df = df.stack().reset_index().rename(columns = {'level_1': 'name'})

    id  name    x   y   z
0   1   a       1   2   3
1   1   b       4   5   6
2   2   a       7   8   9
3   2   b       10  11  12

Answer 3

An addition to the already excellent answers; pivot_longer from pyjanitor can help to abstract the reshaping :

# pip install pyjanitor 
import pandas as pd
import janitor
df.pivot_longer(index = 'id', 
                names_to = ('name', '.value'), 
                names_pattern = r"(.)(.)")
 
   id name   x   y   z
0   1    a   1   2   3
1   2    a   7   8   9
2   1    b   4   5   6
3   2    b  10  11  12