CODEWITHSUNDEEP
pythonarraysnumpynanmean
Jorre Goossens
Jorre Goossens

Reputation: 139

Taking a mean of the elements of arrays with different sizes

Let's say I have a 5x5 array 'A'. I want to take the mean of five elements in that array. It is possible that one of these values is Nan. I thought something like this would work:

np.nanmean(np.array([A[1,1], A[2, 2:3], A[3, 1:3]]))

But it doesn't. I get

ValueError: setting an array element with a sequence.

I also tried concatenating, flattening and using a list instead of np.array, but without luck.

I'm sorry if this question is a duplicate. It seems like an easy problem, but I can't manage to figure it out and I find it hard to pick good search terms to find a solution online.

Upvotes: 0

Views: 156

Answers (1)

jmd_dk
jmd_dk

Reputation: 13100

You can concatenate all the elements into an array before doing the mean:

np.nanmean(np.concatenate([[A[1,1]], A[2, 2:3], A[3, 1:3]]))

Note that I have placed A[1,1] inside an extra list. This is subtle, and the root of your troubles: Though e.g. A[2, 2:3] contains only a single number, it is still an array because it is constructed from a slice. On the other hand, A[1,1] is just a number, not living inside of an array object. Your error message is telling you that mixing this bare number with the other arrays leads to trouble.

Upvotes: 1

Related Questions