Counting elements from one matrix, storing information in another

Question

Hello, I’m a beginner in python and I’m doing an exercise that consists in reading all the numbers of a matrix and adding in another matrix the numbers existing in the previous one and how often they appear. Consequently, the matrix 2 will have two lines, one containing the numbers, and the second how many times they appear. For example, if there is a number 10 and it appears 3 times, the second matrix will have in the first row the number 10, and in the second the number 3, the two in the same column.

NOTE: Consider the first matrix as any already declared.

...

matriz2=[]
linha1 = []
linha2 = []
qntde=0
linha1.append(0)
for l in range(numerodelinhas):
    for c in range(numerodecolunas):
        qntde = 0
        if(linha1.count(matriz[l][c])==0):
            if(l==0 and c == 0):
                linha1[0]=matriz[0][0]
            if(l != 0 or c != 0):
                linha1.append(matriz[l][c])
            for n in range(numerodelinhas):
                qntde = qntde + matriz[n].count(matriz[l][c])
            linha2.append(qntde)

matriz2.append(linha1)
matriz2.append(linha2)
cm2=0
for n in matriz2[0]:
    print (matriz2[0][cm2]," - ",matriz2[1][cm2])
    cm2+=1

The code worked, but I don’t like it at all. There would be a way to make it dry?

Answer 1

You can take advantage of the class Collections. Counter, who receives a iterable or mapping and count the elements. The only adjustment is to turn the matrix into a list, but this is easily solved with the method itertools.chain.().

Follow an example of use:

from itertools import chain
from collections import Counter

# Matrix 4x4 apenas para exemplo
matriz = [
    [2, 4, 6, 8],
    [1, 2, 3, 5],
    [1, 1, 10, 20],
    [1, 2, 4, 7],
]

contador = Counter(chain.from_iterable(matriz))

# printa o objeto Counter, sendo as chaves o elemento
# contado e os valores quantas vezes eles aparecem na matriz
print(contador)

Output:

Counter({
    1: 4,
    2: 3,
    4: 2,
    6: 1,
    8: 1,
    3: 1,
    5: 1,
    10: 1,
    20: 1,
    7: 1
})

Repl.it

Edit: chain(*matriz) changed to chain.from_iterable(matriz) to keep the matrix loading Lazy. (tip: @nosklo)

Answer 2

There is a much easier way to deal with arrays and arrays in python, which is to use a library called numpy. Example:

In [1]: import numpy as np

# matriz1 gerada aleatoriamente de números de 0 a 5 com tamanho 3x3
In [2]: matriz1 = np.random.randint(0, 5, size=(3,3))

In [3]: matriz1
Out[3]: 
array([[1, 2, 4],
       [3, 1, 1],
       [2, 4, 4]])

# retorna os elementos únicos e o número de vezes em que cada um aparece
In [4]: elementos, contagem = np.unique(matriz1, return_counts=True)


# empilha os dois arrays de forma vertical, elementos em cima e contagem embaixo
In [5]: matriz2 = np.vstack([elementos, contagem])

In [6]: matriz2
Out[6]: 
array([[1, 2, 3, 4],
       [3, 2, 1, 3]])

For more information and documentation: http://www.numpy.org. Numpy is the base library for those who work in machine learning and data science. There are several tutorials available, worth checking out.