How to stack several column into one column but keep a column as it is in python-CodePudding

let say i have this dataframe

A	B	C
1	p	q
1		q
2	p
2	p	q

and I want to stack column B and C into one column (let say column D), but keep column A as it is. how to get this? Here the expected output I want

A	D
1	p
1	q
1	q
2	p
2	p
2	q

Note: as far as i know, to stack column B and C into column D i can use this code:

df_temp = df[['B','C']]

df_temp = df_temp.stack().reset_index(name='D')

CodePudding user response：

Use DataFrame.set_index with column A:

df_temp = (df.set_index('A')[['B','C']]
              .stack()
              .reset_index(level=1, drop=True)
              .reset_index(name='D'))
print (df_temp)
   A  D
0  1  p
1  1  q
2  1  q
3  2  p
4  2  p
5  2  q

Or:

df_temp = (df.melt(id_vars='A',value_vars=['B','C'], value_name='D')
             .drop('variable', axis=1)
             .dropna(subset=['D'])
             .sort_values('A', ignore_index=False))
print (df_temp)
   A  D
0  1  p
4  1  q
5  1  q
2  2  p
3  2  p
7  2  q