How to get indices of valid values of numpy masked array

Question

I have a netCDF4 dataset representing multiple matrices of the same dimension (551, 146), one matrix (M1) contains longitude values, another matrix (M2) contains latitude values. Each matrix is a numpy masked array.

Given a lon/lat tuple, (A, B), I want to get the matrix indices (lon, lat) where value A matches in M1 and value B matches in M2.

I thought I could represent the indices with:

lon_idx, lat_idx = np.mgrid[:lon.shape[0], :lon.shape[1]]

and two matrices of same shape where one have all values set to A and the other have all values set to B.

Then I am hoping to somehow combine these matrices and end up with an array of lon, lat indices where the values matched.

What is the idiomatic way to do this in numpy?

Please include a [MCVE], including the output you're looking for. — sacuL, Mar 22 '18 at 19:36
@Divakar I think I would need more than that hint to understand how to solve my problem. I've tried using `np.ix_`, but it takes 1-d arrays, and I have matrices, i.e 2d arrays.. — jollyroger, Mar 22 '18 at 19:54
Try printing `np.where(M1 == A)` and see what you get. You should get the indices you need. — jmd_dk, Mar 22 '18 at 20:51

score 2 · Answer 1 · answered Mar 22 '18 at 20:43

2

Thanks to SnoopJeDi on #python @freenode who found the solution for me:

In [58]: np.argwhere((lon == lon[250][145]) & (lat == lat[250][145]))
Out[58]: array([[250, 145]])

answered Mar 22 '18 at 20:43

jollyroger

659
1
10
19

How to get indices of valid values of numpy masked array

1 Answers1