"TypeError: 'DataFrame' objects are mutable, thus they cannot be hashed" while sorting pandas dataframe index

Question

I have a following dataframe h:

In [24]: h.head()
Out[24]: 
                 alpha1  alpha2    gamma1  gamma2       chi2min gender  age
filename                                                                   
F35_HC_532d.dat  0.0000   0.000       NaN    0.00  1.000000e+25      F   35
M48_HC_551d.dat  0.7353   3.943  0.425922    0.15  2.072617e+01      M   48
M24_HC_458d.dat  0.7777   4.754  0.463753    0.15  1.390893e+01      M   24
M48_HC_552d.dat  0.7633   3.672  0.394370    0.15  1.965052e+01      M   48
M40_HC_506d.dat  0.7793   3.271  0.513597    0.20  1.089716e+01      M   40

I am trying to sort the dataframe index according to age values:

In [25]: h.sort_index(h.sort_values('age'))

This throws an error:

TypeError: 'DataFrame' objects are mutable, thus they cannot be hashed

What am I missing? Any ideas?

Solution 1

Is that what you want?

In [14]: h
Out[14]:
                 alpha1  alpha2    gamma1  gamma2       chi2min gender  age
filename
F35_HC_532d.dat  0.0000   0.000       NaN    0.00  1.000000e+25      F   35
M48_HC_551d.dat  0.7353   3.943  0.425922    0.15  2.072617e+01      M   48
M24_HC_458d.dat  0.7777   4.754  0.463753    0.15  1.390893e+01      M   24
M48_HC_552d.dat  0.7633   3.672  0.394370    0.15  1.965052e+01      M   48
M40_HC_506d.dat  0.7793   3.271  0.513597    0.20  1.089716e+01      M   40

In [15]: h.sort_values('age')
Out[15]:
                 alpha1  alpha2    gamma1  gamma2       chi2min gender  age
filename
M24_HC_458d.dat  0.7777   4.754  0.463753    0.15  1.390893e+01      M   24
F35_HC_532d.dat  0.0000   0.000       NaN    0.00  1.000000e+25      F   35
M40_HC_506d.dat  0.7793   3.271  0.513597    0.20  1.089716e+01      M   40
M48_HC_551d.dat  0.7353   3.943  0.425922    0.15  2.072617e+01      M   48
M48_HC_552d.dat  0.7633   3.672  0.394370    0.15  1.965052e+01      M   48

Solution 2

I think your index is filename. Maybe you could try something like:

h['index1'] = h.index
h.sort_values(by=['index1', 'age'])

But also it will not make so much sense since it will not change the order. Alternatively you can try:

h.sort_values(by='age')

Then:

h.reindex([range(some_number)])

Comments

[MaxU - stop genocide of UA]

Are you after: h.sort_values('age')? Can you provide a desired data set?

[Peaceful]

@MaxU : I think I didn't get you. I am trying to arrange the rows of the dataframe according to the age values.

[Peaceful]

That's another method. I want to know what is wrong with my method.

[Peaceful]

I realized now. That's right. I was making a silly mistake.

[Cenk_Mitir]

the problem in your method is indices are filenames not numbers