How to melt 2 columns at the same time?
Question:
In Pandas, I have the following data frame:
id1 id2 t1 l1 t2 l2
0 1 2 a b c d
1 3 4 g h i j
I would like to melt two columns at once. That is, the desired output is:
id1 id2 tz lz
0 1 2 a b
1 1 2 c d
2 3 4 g h
3 3 4 i j
I know standard melting:
d.melt(id_vars=['id1', 'id2'],
value_vars=['t1', 't2', 'l1', 'l2'])
but that stacks all columns
id1 id2 variable value
0 1 2 t1 a
1 3 4 t1 g
2 1 2 t2 c
3 3 4 t2 i
4 1 2 l1 b
5 3 4 l1 h
6 1 2 l2 d
7 3 4 l2 j
How could I melt two columns at once? Something like:
d.melt(id_vars=['id1', 'id2'],
value_vars={('t1', 'l1'): 'tz', ('t2', 'l2'): 'lz'})
would be great.
Answers:
This is wide_to_long
pd.wide_to_long(df, stubnames=['t','l'], i=['id1','id2'], j='drop').reset_index(level=[0,1])
Out[52]:
id1 id2 t l
drop
1 1 2 a b
2 1 2 c d
1 3 4 g h
2 3 4 i j
You can use melt
twice here and after that concat them to get desired output:
t = d.melt(id_vars=['id1', 'id2'], value_vars=['t1', 't2'], value_name='tz').drop('variable', axis=1)
l = d.melt(id_vars=['id1', 'id2'], value_vars=['l1', 'l2'], value_name='lz').iloc[:, -1:]
df = pd.concat([t, l], axis=1).sort_values('id1')
Output
print(df)
id1 id2 tz lz
0 1 2 a b
2 1 2 c d
1 3 4 g h
3 3 4 i j
In Pandas, I have the following data frame:
id1 id2 t1 l1 t2 l2
0 1 2 a b c d
1 3 4 g h i j
I would like to melt two columns at once. That is, the desired output is:
id1 id2 tz lz
0 1 2 a b
1 1 2 c d
2 3 4 g h
3 3 4 i j
I know standard melting:
d.melt(id_vars=['id1', 'id2'],
value_vars=['t1', 't2', 'l1', 'l2'])
but that stacks all columns
id1 id2 variable value
0 1 2 t1 a
1 3 4 t1 g
2 1 2 t2 c
3 3 4 t2 i
4 1 2 l1 b
5 3 4 l1 h
6 1 2 l2 d
7 3 4 l2 j
How could I melt two columns at once? Something like:
d.melt(id_vars=['id1', 'id2'],
value_vars={('t1', 'l1'): 'tz', ('t2', 'l2'): 'lz'})
would be great.
This is wide_to_long
pd.wide_to_long(df, stubnames=['t','l'], i=['id1','id2'], j='drop').reset_index(level=[0,1])
Out[52]:
id1 id2 t l
drop
1 1 2 a b
2 1 2 c d
1 3 4 g h
2 3 4 i j
You can use melt
twice here and after that concat them to get desired output:
t = d.melt(id_vars=['id1', 'id2'], value_vars=['t1', 't2'], value_name='tz').drop('variable', axis=1)
l = d.melt(id_vars=['id1', 'id2'], value_vars=['l1', 'l2'], value_name='lz').iloc[:, -1:]
df = pd.concat([t, l], axis=1).sort_values('id1')
Output
print(df)
id1 id2 tz lz
0 1 2 a b
2 1 2 c d
1 3 4 g h
3 3 4 i j