CODEWITHSUNDEEP
arraysnumpy
Stoddard
Stoddard

Reputation: 55

How do I find the conditional intersection of different elements in 2D numpy array

I have an array that has unique pairs of integers that is fully sorted like this:

x = np.array([[40, 202], [44, 117], [44, 125], [45, 53], [45, 675], [45, 1013], [53, 675], [116, 845], [117, 125], [126, 130], [130, 144]])

How can I obtain values of a, b, c that occur in three different tuples [a, b] and [a, c] and [b, c]? For example, [44, 117] and [44, 125] and [117, 125] fulfill this pattern with a = 44, b = 117, and c = 125.

Looking at it vertically is perhaps an easier way to visualize what I’m looking for. T_1, T_2, and T_3 refer to individual tuples in array x. 

 T_1 = [a, b] eg [44, 117]  
 T_2 = [a, c] eg [44, 125]  
 T_3 = [b, c] eg [117, 125]

Simulated code and output would be:
def findTriple(arr): """Returns a 2darray with tuple a, b, c for each match that results in the union [a, b] and [a, c] and [b, c] in array x""" # do something return res data = findTriple(x) print (data) #two answers for example x given.

[[44, 117, 125], [45, 53, 675]]

I couldn't figure out a way to create a mask or use list comprehension.

Upvotes: 1

Views: 41

Answers (1)

Mr Tarsa
Mr Tarsa

Reputation: 6652

You can accomplish it with list comprehension

>> [(x1, x2, xx2) for x1, x2 in x for xx1, xx2 in x if x2 == xx1]
[(44, 117, 125), (45, 53, 675), (126, 130, 144)]

Upvotes: 1

Related Questions