CODEWITHSUNDEEP
pythonlist
Ben Jones
Ben Jones

Reputation: 565

Get n highest y values from list of coordinate pairs in Python 3

I'm aware of how to get the n highest values in a list but I want a quick function that gets the highest y values of a coordinate pair.

I know the following gives the n highest values in a list 

[a[i] for i in np.argsort(a)[-n:]]

But what I really want is to feed it coordinates of the form

[[1,2], [3,4], [6,2], [6,11], [1,5]]

and get n coordinates with the highest y values from it, i.e. for n = 2

[[6,11], [1,5]]

But I can't quite get past the last step. Not having to use external libraries other than numpy would be a bonus. I tried changing the axis of argsort(list, axis=0) but that didn't work. Thanks

Upvotes: 3

Views: 760

Answers (4)

Divakar
Divakar

Reputation: 221614

Here's one NumPy based approach with np.argsort -

n = 2 # Number of entries to keep
arr = np.asarray(input_list) # Convert to array for further processing
out = arr[arr[:,1].argsort()[-n:]].tolist()

Another NumPy based one and should be faster with np.argpartition -

out = arr[(-arr[:,1]).argpartition(n,axis=0)[:n]].tolist()

Upvotes: 1

Pythonista
Pythonista

Reputation: 11635

As an alternative, terser, answer you could use nlargest from heapq

nlargest(n, your2DList, key=lambda x: x[-1])

Upvotes: 4

Jean-François Fabre
Jean-François Fabre

Reputation: 140216

just sort the list using 2nd coord as key, and print the 2 last items:

z = [[1,2], [3,4], [6,2], [6,11], [1,5]]

print(sorted(z,key=lambda x:x[1])[-2:])

result:

[[1, 5], [6, 11]]

Upvotes: 2

Alexey Guseynov
Alexey Guseynov

Reputation: 5304

A sorted function takes named argument key which is used to compute a compare key for each value.

>>> sorted([[1,2], [3,4], [6,2], [6,11], [1,5]], key=lambda pair: pair[1])[-n:]
[[1, 5], [6, 11]]

Upvotes: 3

Related Questions