CODEWITHSUNDEEP
pythonpandasdataframesorting
user16470918
user16470918

Reputation: 67

How to sort values in pandas based on a column that has repeating values

I have the following dataframe:

           name_1  name_2   Temp      x1   x2  
0           a        b     293.35     0.0  1.0  
0           a        b     293.35     0.5  0.5  
0           a        b     293.35     1.0  0.0  
0           a        b     295.35     0.0  1.0  
0           a        b     295.35     0.5  0.5  
0           a        b     295.35     1.0  0.0  
0           a        b     297.35     0.0  1.0  
0           a        b     297.35     0.5  0.5  
0           a        b     297.35     1.0  0.0 
1           c        d     294.35     0.0  1.0  
1           c        d     294.35     0.5  0.5  
1           c        d     294.35     1.0  0.0  
1           c        d     296.35     0.0  1.0  
1           c        d     296.35     0.5  0.5  
1           c        d     296.35     1.0  0.0  
1           c        d     299.35     0.0  1.0  
1           c        d     299.35     0.5  0.5  
1           c        d     299.35     1.0  0.0
2           a        c     300.35     0.0  1.0  
2           a        c     305.35     0.5  0.5  
2           a        c     310.35     1.0  0.0

I would like to arrange the data such that the Temp are repeating based on x1 and x2 to get this:

           name_1 name_2   Temp       x1   x2  
0           a        b     293.35    0.0  1.0 
0           a        b     295.35    0.0  1.0  
0           a        b     297.35    0.0  1.0  
0           a        b     293.35    0.5  0.5  
0           a        b     295.35    0.5  0.5  
0           a        b     297.35    0.5  0.5 
0           a        b     293.35    1.0  0.0  
0           a        b     295.35    1.0  0.0  
0           a        b     297.35    1.0  0.0
1           c        d     294.35    0.0  1.0  
1           c        d     296.35    0.0  1.0  
1           c        d     299.35    0.0  1.0  
1           c        d     294.35    0.5  0.5  
1           c        d     296.35    0.5  0.5  
1           c        d     299.35    0.5  0.5  
1           c        d     294.35    1.0  0.0  
1           c        d     296.35    1.0  0.0  
1           c        d     299.35    1.0  0.0
2           a        c     300.35    0.0  1.0  
2           a        c     305.35    0.5  0.5  
2           a        c     310.35    1.0  0.0

I got close using the following:

df['new'] = df.groupby('Temp')['Temp'].cumcount()
df.sort_values(by=['new', 'Temp', "name_1", "name_2"])

This code yields:

           name_1 name_2    Temp       x1   x2  
0           a        b     293.35     0.0  1.0 
1           c        d     294.35     0.0  1.0  
0           a        b     295.35     0.0  1.0  
2           a        c     300.35     0.0  1.0

The issue with this is that TRange is not being repeated based on name_1 and name_2 like the desired output.

Upvotes: 0

Views: 907

Answers (2)

Corralien
Corralien

Reputation: 120419

Sort by x1 values then by index using kind='mergesort' to ensure your values stay sorted:

>>> df.sort_values('x1', kind='mergesort').sort_index(kind='mergesort')

  name_1 name_2    Temp   x1   x2
0      a      b  293.35  0.0  1.0
0      a      b  295.35  0.0  1.0
0      a      b  297.35  0.0  1.0
0      a      b  293.35  0.5  0.5
0      a      b  295.35  0.5  0.5
0      a      b  297.35  0.5  0.5
0      a      b  293.35  1.0  0.0
0      a      b  295.35  1.0  0.0
0      a      b  297.35  1.0  0.0
1      c      d  294.35  0.0  1.0
1      c      d  296.35  0.0  1.0
1      c      d  299.35  0.0  1.0
1      c      d  294.35  0.5  0.5
1      c      d  296.35  0.5  0.5
1      c      d  299.35  0.5  0.5
1      c      d  294.35  1.0  0.0
1      c      d  296.35  1.0  0.0
1      c      d  299.35  1.0  0.0
2      a      c  300.35  0.0  1.0
2      a      c  305.35  0.5  0.5
2      a      c  310.35  1.0  0.0

Upvotes: 1

mozway
mozway

Reputation: 260745

Simply sort on x1. You can specify a stable sorting algorithm:

df = df.sort_values(by='x1', kind='mergesort')

grouping and applying a cumcount is going nothing other than keeping the original order of the other parameters, which a stable sorting algorithm is already doing.

cf. sort_values documentation for reference.

Upvotes: 1

Related Questions