Sorting arrays in NumPy by column

Question

How can I sort an array in NumPy by the nth column   For example   a   array   9  2  3               4  5  6               7  0  5      I d like to sort rows by the second column  such that I get back   array   7  0  5           9  2  3           4  5  6

User · Answer

Here is another solution considering all columns  more compact way of J J s answer    ar np array   0  0  0  1                 1  0  1  0                 0  1  0  0                 1  0  0  1                 0  0  1  0                 1  1  0  0      Sort with lexsort   ar np lexsort   ar    i  for i in range ar shape 1 -1  -1  -1        Output   array   0  0  0  1           0  0  1  0           0  1  0  0           1  0  0  1           1  0  1  0           1  1  0  0

User · Answer

steve s answer is actually the most elegant way of doing it   For the  correct  way see the order keyword argument of numpy ndarray sort   However  you ll need to view your array as an array with fields  a structured array    The  correct  way is quite ugly if you didn t initially define your array with fields     As a quick example  to sort it and return a copy   In  1   import numpy as np  In  2   a   np array   1 2 3   4 5 6   0 0 1     In  3   np sort a view  i8 i8 i8    order   f1    axis 0  view np int  Out 3    array   0  0  1           1  2  3           4  5  6      To sort it in-place   In  6   a view  i8 i8 i8   sort order   f1    axis 0    lt -- returns None  In  7   a Out 7    array   0  0  1           1  2  3           4  5  6       Steve s really is the most elegant way to do it  as far as I know      The only advantage to this method is that the  order  argument is a list of the fields to order the search by  For example  you can sort by the second column  then the third column  then the first column by supplying order   f1   f2   f0

User · Answer

You can sort on multiple columns as per Steve Tjoa s method by using a stable sort like mergesort and sorting the indices from the least significant to the most significant columns   a   a a   2  argsort      First sort doesn t need to be stable  a   a a   1  argsort kind  mergesort    a   a a   0  argsort kind  mergesort      This sorts by column 0  then 1  then 2

User · Answer

It is an old question but if you need to generalize this to a higher than 2 dimension arrays  here is the solution than can be easily generalized     np einsum  ij- gt ij   a a   1  argsort         This is an overkill for two dimensions and a a   1  argsort    would be enough per  steve s answer  however that answer cannot be generalized to higher dimensions  You can find an example of 3D array in this question   Output     7 0 5    9 2 3    4 5 6

User · Answer

for sorting along column 1 indexofsort np argsort dataset   0  axis -1 kind  stable    dataset     dataset indexofsort

User · Answer

From the Python documentation wiki  I think you can do   a      1  2  3    4  5  6    0  0  1      a   sorted a  key lambda a entry  a entry 1    print a   The output is      0  0  1    1  2  3    4  5  6

User · Answer

I had a similar problem   My Problem   I want to calculate an SVD and need to sort my eigenvalues in descending order  But I want to keep the mapping between eigenvalues and eigenvectors  My eigenvalues were in the first row and the corresponding eigenvector below it in the same column   So I want to sort a two-dimensional array column-wise by the first row in descending order   My Solution  a   a     a 0   argsort     -1     So how does this work   a 0   is just the first row I want to sort by   Now I use argsort to get the order of indices   I use    -1  because I need descending order   Lastly I use a          to get a view with the columns in the right order

User · Answer

import numpy as np a np array   21 20 19 18 17   16 15 14 13 12   11 10 9 8 7   6 5 4 3 2    y np argsort a   2  kind  mergesort    a   2   19 14 9 4  a a y  print a   Desired output is   6 5 4 3 2   11 10 9 8 7   16 15 14 13 12   21 20 19 18 17   note that argsort numArray  returns the indices of an numArray as it was supposed to be arranged in a sorted manner  example x np array  8 1 5    z np argsort x    1 3 0  are the   indices of the predicted sorted array   print x z    boolean indexing which sorts the array on basis of indices saved in z  answer would be  1 5 8

User · Answer

A little more complicated lexsort example - descending on the 1st column  secondarily ascending on the 2nd   The tricks with lexsort are that it sorts on rows  hence the  T   and gives priority to the last   In  120   b np array   1 2 1   3 1 2   1 1 3   2 3 4   3 2 5   2 1 6    In  121   b Out 121    array   1  2  1           3  1  2           1  1  3           2  3  4           3  2  5           2  1  6    In  122   b np lexsort   1 -1  b    1 0    T   Out 122    array   3  1  2           3  2  5           2  1  6           2  3  4           1  1  3           1  2  1

User · Answer

From the NumPy mailing list  here s another solution    gt  gt  gt  a array   1  2           0  0           1  0           0  2           2  1           1  0           1  0           0  0           1  0          2  2     gt  gt  gt  a np lexsort np fliplr a  T   array   0  0           0  0           0  2           1  0           1  0           1  0           1  0           1  2           2  1           2  2

User · Answer

def sort np array x  column None  flip False       x   x np argsort x    column        if flip          x   np flip x  axis 0      return x  Array in the original question  a   np array   9  2  3                  4  5  6                  7  0  5     The result of the sort np array function as expected by the author of the question  sort np array a  column 1  flip False    2   array   7  0  5                9  2  3                4  5  6

User · Answer

I suppose this works  a a   1  argsort     This indicates the second column of a and sort it based on it accordingly

User · Answer

In case someone wants to make use of sorting at a critical part of their programs here s a performance comparison for the different proposals   import numpy as np table   np random rand 5000  10    timeit table view  f8 f8 f8 f8 f8 f8 f8 f8 f8 f8   sort order   f9    axis 0  1000 loops  best of 3  1 88 ms per loop   timeit table table   9  argsort    10000 loops  best of 3  180   s per loop  import pandas as pd df   pd DataFrame table   timeit df sort values 9  ascending True  1000 loops  best of 3  400   s per loop   So  it looks like indexing with argsort is the quickest method so far

User · Answer

Simply using sort  use coloumn number based on which you want to sort   a   np array  1 1    1 -1    -1 1    -1 -1    print  a  a a tolist    a   np array sorted a  key lambda a entry  a entry 0    print  a

[python] Sorting arrays in NumPy by column

Examples related to python

Examples related to arrays

Examples related to sorting

Examples related to numpy

Examples related to scipy