Set values in xarray DataArray to NaN if all values across a dimension are zero

Question

I have an xarray like this:

import xarray as xr

da1 = xr.DataArray([[0, 1, 5, 5], [1, 2, 2, 0], [9, 3, 2, 0]], dims=['x', 'y'])
da2 = xr.DataArray([[0, 2, 9, 3], [0, 0, 7, 0], [0, 2, 6, 0]], dims=['x', 'y'])
da3 = xr.DataArray([[0, 7, 2, 0], [7, 2, 6, 0], [0, 6, 1, 0]], dims=['x', 'y'])

combined = xr.concat([da1, da2, da3], 'band')

It looks like this:

array([[[0, 1, 5, 5],
        [1, 2, 2, 0],
        [9, 3, 2, 0]],

       [[0, 2, 9, 3],
        [0, 0, 7, 0],
        [0, 2, 6, 0]],

       [[0, 7, 2, 0],
        [7, 2, 6, 0],
        [0, 6, 1, 0]]])

with three dimensions: band, x and y.

I want to set the values in this array to NaN in the situations where all the values across the band dimension are zero. For example, the values at combined.isel(x=0, y=0) should be set to NaN, as all those values are zero, but the values at combined.isel(x=1, y=1) should not be, as only one of the values is zero.

How can I do this?

I've tried using:

combined.where(combined != 0)

but this sets all values that are zero to NaN, which doesn't do what I want.

I then tried something like:

combined.where((combined.isel(band=0) != 0) & (combined.isel(band=1) != 0) & (combined.isel(band=2) != 0))

but the 'and' bit doesn't seem to work properly and it gives a strange (and incorrect) result.

Update: As an extension, I would like to be able to do the same thing but for very small values, rather than zeros. For example, setting all values across the band dimension to NaN if all values across that dimension are < 0.01. Is there an easy way to do this?

Any advice very much appreciated

Thrasy · Accepted Answer · 2022-01-10 10:40:54Z

1

You can do that with:

combined.where(combined.any(dim = 'band'))

answered Jan 10, 2022 at 10:40

Thrasy

6065 silver badges9 bronze badges

Sign up to request clarification or add additional context in comments.

3 Comments

robintw Over a year ago

Thank you. Having tried this, I've actually realised I need to do something slightly more complicated - I've updated the question, would you be able to take a quick look?

Thrasy Over a year ago

@robintw It's certainly not optimal but I think you can do it with clip : (combined.clip(min = 0, max = cutoff) != combined).any(dim = "band")

Thrasy Over a year ago

@robintw Glad it helped. By the way, your free GIS data list is an old favourite of mine. Very handy!

Collectives™ on Stack Overflow

Set values in xarray DataArray to NaN if all values across a dimension are zero

1 Answer 1

3 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

3 Comments

Your Answer

Sign up or log in

Post as a guest

Related