CODEWITHSUNDEEP
pythonpython-3.xpandassorting
Acinonyx
Acinonyx

Reputation: 89

Sorting columns of pandas on every row

I am trying to sort the columns of a panda on every row. Consider this dataframe df:

X C1 C2 C3 C4 Y Z
A 11 15 12 13 A1 A2
B 21 25 22 23 B1 B2
C 31 35 32 33 C1 C2

I am trying to sort columns in ascending order C1-C4 on every row such that the end dataframe looks like this:

X C1 C2 C3 C4 Y Z
A 11 12 13 15 A1 A2
B 21 22 23 25 B1 B2
C 31 32 33 35 C1 C2

I have looked up pandas sort_values() but having difficulty with the manual. Hoping somebody can show me some tricks. I am trying to sort only a few columns and not all on each row.

Upvotes: 2

Views: 84

Answers (2)

jezrael
jezrael

Reputation: 862521

Use numpy.sort for sorting by all rows:

cols = ['C1','C2','C3','C4']
df[cols] = np.sort(df[cols], axis=1)
print (df)
   X  C1  C2  C3  C4   Y   Z
0  A  11  12  13  15  A1  A2
1  B  21  22  23  25  B1  B2
2  C  31  32  33  35  C1  C2

If possible, sort by index 0:

cols = ['C1','C2','C3','C4']
df[cols] = df[cols].sort_values(0, axis=1)
#thanks @pygo for another solution
df[cols].sort_values(0, axis=1, inplace=True)
print (df)
   X  C1  C2  C3  C4   Y   Z
0  A  11  12  13  15  A1  A2
1  B  21  22  23  25  B1  B2
2  C  31  32  33  35  C1  C2

Difference between solutions with changed input DataFrame:

print (df)
   X  C1  C2  C3  C4   Y   Z
0  A  11  15  12  13  A1  A2
1  B   2   1   5   4  B1  B2
2  C  31  35  32  33  C1  C2

cols = ['C1','C2','C3','C4']
df[cols] = np.sort(df[cols], axis=1)
print (df)
   X  C1  C2  C3  C4   Y   Z
0  A  11  12  13  15  A1  A2
1  B   1   2   4   5  B1  B2
2  C  31  32  33  35  C1  C2

cols = ['C1','C2','C3','C4']
df[cols] = df[cols].sort_values(0, axis=1)
print (df)
   X  C1  C2  C3  C4   Y   Z
0  A  11  12  13  15  A1  A2
1  B   2   5   4   1  B1  B2
2  C  31  32  33  35  C1  C2

Upvotes: 3

Mayank Porwal
Mayank Porwal

Reputation: 34046

You can use pandas function sort_values() like below: 

In [331]: df
Out[331]: 
   X  C1  C2  C3  C4    Y Z
0  A  11  15  12  13  A1 A2
1  B  21  25  22  23  B1 B2
2  C  31  35  32  33  C1 C2

In [332]: df.sort_values(['C1','C2','C3','C4'])
Out[332]: 
   X  C1  C2  C3  C4    Y Z
0  A  11  15  12  13  A1 A2
1  B  21  25  22  23  B1 B2
2  C  31  35  32  33  C1 C2

Upvotes: 2

Related Questions