相同数组的两个不同列的相加

Question

I have an array which hold an integer value and item Id value in it.

I want to add two different values of two different items. For example, let's say foretold item is like this:

array1 = [[12, 1], [23, 2], [34, 3]]

and the desired output should be something like that:

[[35, 12, 1, 23, 2], [46, 12, 1, 34, 3], [57, 23, 2, 34, 3]]

For small array counts there is no problem but when the count of array exceed a hundred I encounter with huge performance problems.

Is there any method for this in Python?

My example code:

for item1 in array1:
    for item2 in array1:
        sumTwoOutput.append([item1[0] + item2[0], item1[0], item1[1], item2[0], item2[1]])

dfSumTwoOutput = pd.DataFrame(sumTwoOutput)
dfSumTwoOutput.columns = ["OutputSum", "Value1", "ItemId1", "Value2", "ItemId2"]

Answer 1

You can use itertools.combinations :

from itertools import combinations

array = [[12, 1], [23, 2], [34, 3]]

sumTwoOutput = []
for (num1, id1), (num2, id2) in combinations(array, 2):
    sumTwoOutput.append([num1 + num2, num1, id1, num2, id2])

print(sumTwoOutput)

This will give:

[[35, 12, 1, 23, 2], [46, 12, 1, 34, 3], [57, 23, 2, 34, 3]]

---

As I see from your code that you are using this with pandas , here is another way relying more on pandas :

import pandas as pd
from itertools import combinations

array = [[12, 1], [23, 2], [34, 3]]

df = pd.DataFrame((*x, *y) for x,y in combinations(array, 2))
df.columns = ["Value1", "ItemId1", "Value2", "ItemId2"]

df.insert(0, "Sum", df["Value1"] + df["Value2"])

print(df)

Gives:

   Sum  Value1  ItemId1  Value2  ItemId2
0   35      12        1      23        2
1   46      12        1      34        3
2   57      23        2      34        3

Answer 2

You can use itertools package.

from itertools import combinations

array = [[12, 1], [23, 2], [34, 3]]

sumTwoOutput = []
comb = combinations(range(len(array)), 2)

for i, j in comb:
    sumTwoOutput.append([array[i][0] + array[j][0], array[i][0], array[i][1], array[j][0], array[j][1]])

print(sumTwoOutput)
# [[35, 12, 1, 23, 2], [46, 12, 1, 34, 3], [57, 23, 2, 34, 3]]