Loop through dictionary to create a dataframe from function without repeating keys

Question

dict = {'A': ndarray_1, 'B': ndarray_2, 'C': ndarray_3}

I'm trying, without success so far, to use a for loop it in a callable function of a package that return two results and make a dataframe like this:

Letter | Letter |             Result 1            |          Result 2

  A    |    B   |    result1(ndarray_1,ndarray_2) |  result2(ndarray_1,ndarray_2)
  A    |    C   |    result1(ndarray_1,ndarray_3) |  result2(ndarray_1,ndarray_3)
  B    |    C   |    result1(ndarray_2,ndarray_3) |  result2(ndarray_2,ndarray_3)

I'm having trouble acessing the value of each key to use in the function and not repeating ({A,B}, {B,A}, for example)

Answer 1

There might be built-in functions to do this, but here's a simple way to do it from scratch:

1. Create an ordered list of the keys: keys = dict.keys()
2. Loop through the indices of the ordered list: for i in range(len(keys)):
3. Inside that, loop through the indices starting at i+1: for j in range(i + 1, len(keys)):
4. Inside that, add whatever results to your dataframe using the keys at i and j.

All together:

keys = dict.keys()

for i in range(len(keys)):
    for j in range(i + 1, len(keys)):
        # add a new row to the dataframe using keys[i] and keys[j]

For the intuition, imagine taking a matrix of all dividing it in half diagonally, and then iterating through only the top half: