Sort all matrix

I'm trying to sort a matrix if I have

[[5 7 1 2]
 [4 8 1 4] 
 [4 0 1 9]
 [2 7 5 0]]

I want this result

[[9 8 7 7]
 [5 5 4 4]
 [4 2 2 1]
 [1 1 0 0]]

this is my code, I'm using random numbers

import numpy as np

MatriX = np.random.randint(10, size=(4,4))
print(MatriX)
for i in range(10):
    for j in range(10):
        a=np.sort(-MatriX,axis=1)*-1
        b=np.sort(-a,axis=0)*-1
        c=np.transpose(b)
   
        d=np.sort(-c,axis=1)*-1
        e=np.sort(-d,axis=0)*-1
 
    
print(e)

but this code, don't work at all

Solution

It's as simple as this:

import numpy as np

data = np.random.randint(10, size=(4,4))
print(data)

result = np.sort(data.flatten())[::-1].reshape(data.shape)

print(result)

Output:

[[4 7 7 6]
 [1 9 5 0]
 [5 8 1 2]
 [2 5 7 3]]
[[9 8 7 7]
 [7 6 5 5]
 [5 4 3 2]
 [2 1 1 0]]

An explanation of the parts of np.sort(data.flatten())[::-1].reshape(data.shape):

data.flatten() flattens the array into a one-dimensional shape and returns the result
np.sort() sorts the one-dimensional result and returns the sorted data
[::-1] reverses the order of the sorted data
.reshape(data.shape) applies the shape of the original data to that reversed sorted data

Note that something like this would be easier to read:

result = data.flattened().sorted().reversed().reshaped(data.shape)

But numpy doesn't implement all the needed methods that allow for this more readable chaining and many methods make changes in-place. So you get the slightly harder to read solution above instead. (you could define a class that has these, but your code wouldn't make sense to others familiar with numpy, so it is probably best left the way it is)

However, if your data is large, performing some operations in-place may make more sense. The solution given here is intended to be simple, you may want to look at some optimisation for very large data.