Sorting arrays in Python by a non-integer column

Question

So for example I have an array that I want to sort by a column in an ascending order, and it's easy to do for integers using 'sorting()', 'np.arrange()', or 'np.argsort()'.

However, what if my column is consisting of floats?

What would you recommend?

Edit: I mean, I have something like:

a = array([[1.7, 2, 3],
           [4.5, 5, 6],
           [0.1, 0, 1]])

and I want to get this:

array([[0.1, 0, 1],
       [1.7, 2, 3],
       [4.5, 5, 6]])

So far with argsort() I get the following error: Type Error: only integer scalar arrays can be converted to a scalar index

Answer 1

You can use a standard Python's sorted (or sort for in-place sorting), no matter what is contained in the sequence. Just use a custom key , or a custom compare function ( cmp ). For example, to sort a list of lists (2-d array) ascending by 4th column:

>>> a=[[1.0,2.0,3.0,4.0], [4.0,3.0,2.0,1.0], [0,0,0,0]]
>>> from operator import itemgetter
>>>> sorted(a, key=itemgetter(3))
[[0, 0, 0, 0], [4.0, 3.0, 2.0, 1.0], [1.0, 2.0, 3.0, 4.0]]

Answer 2

The standard way to do this in numpy is to specify the correct axis you want to sort on, by default it sorts on axis=-1 :

>>> np.sort(a, axis=0)
array([[ 0.1,  0. ,  1. ],
       [ 1.7,  2. ,  3. ],
       [ 4.5,  5. ,  6. ]])

Or inplace:

>>> a.sort(axis=0)
>>> a
array([[ 0.1,  0. ,  1. ],
       [ 1.7,  2. ,  3. ],
       [ 4.5,  5. ,  6. ]])

To sort just on a specific column you can use argsort() , eg column 0 :

>>> a[np.argsort(a[:,0])]
array([[ 0.1,  0. ,  1. ],
       [ 1.7,  2. ,  3. ],
       [ 4.5,  5. ,  6. ]])