I have two arrays
a = array([1,2,3])
b = array([2,7])
Now I want to check if elements of a are in b and the returning answer should be (False, True, False). Is there some simple way to do this without using functions?
4 Answers
How about this:
>>> numpy.setmember1d(a, b)
array([False, True, False], dtype=bool)
update, thanks seberg. With newer verions of numpy this is:
>>> numpy.in1d(a, b)
array([False, True, False], dtype=bool)
1 Comment
seberg
On newer numpy versions the function is called
np.in1d (or np.lib.arraysetops.in1d, but the best solution with arrays, at least if b is not small.With standard python lists:
>>> a = [1, 2, 3]
>>> b = [2, 7]
>>> tuple(x in b for x in a)
(False, True, False)
Assuming that your array function returns an object that also supports both iterations and the in operator, it should work the same.
5 Comments
Bi Rico
This will be quite slow if a and b are large.
poke
@BiRico All valid solutions, without knowing about numpy arrays, are that though :P
Bi Rico
If b is large, I would probably opt for
[x in set(b) for x in a] to avoid looping over b.poke
That will call
set(b) len(a) times though. But yeah, obviously you could make changes to improve the performance.Bi Rico
Yes you are absolutely right, how about
b = set(b); tuple(x in b for x in a), does that work?Using only numpy:
>>> (a[:,None] == b).any(axis=-1)
(So, we transform a from a (N,) to a (N,1) array, then test for equality using numpy's broadcasting. We end up with a (N, M) array (assuming that b had a shape (M,)...), and we just check whether ther's a True on each row with any(axis=-1).
Comments
Well, this is how I'd do it with lists:
>>> a = [1, 2, 3]
>>> b = [2, 7]
>>> result = []
>>>
>>> for x in a:
... result.append(x in b)
...
>>> print result
[False, True, False]
2 Comments
Jakob Bowyer
As other answers have shown this can be crushed down into a list or generator comprehension
Hubro
I started typing my answer before there were any others. This answer may be unnecessarily verbose but it's still valid and could be of help to those who don't yet fully comprehend list comprehensions.
arrayfunction from? numpy?