Best way to reverse along a dimension with large dask data?

[fbunt]

I am working with some gigantic rasters and need to flip them along the y dimension. So far I've been using a method I found on stackoverflow. This works perfectly but for large rasters I get a warning that slicing is producing a large chunk:

import dask.array as da
import xarray as xr

n = 100_000
data = da.zeros((1, n, n), chunks=(1, "auto", "auto"))
x = xr.DataArray(data, dims=("band", "y", "x"), coords=([1], np.arange(n), np.arange(n)))
print(x)
x = x.reindex(y=x.y[::-1])

<xarray.DataArray 'zeros_like-74c41486ec6ed801df96ced5fd27a2aa' (band: 1,
                                                                 y: 100000,
                                                                 x: 100000)>
dask.array<zeros_like, shape=(1, 100000, 100000), dtype=float64, chunksize=(1, 4000, 4000), chunktype=numpy.ndarray>
Coordinates:
  * band     (band) int64 1
  * y        (y) int64 0 1 2 3 4 5 6 ... 99994 99995 99996 99997 99998 99999
  * x        (x) int64 0 1 2 3 4 5 6 ... 99994 99995 99996 99997 99998 99999

/home/USER/anaconda3/envs/main/lib/python3.8/site-packages/xarray/core/indexing.py:1228: PerformanceWarning: Slicing is producing a large chunk. To accept the large
chunk and silence this warning, set the option
    >>> with dask.config.set(**{'array.slicing.split_large_chunks': False}):
    ...     array[indexer]

To avoid creating the large chunks, set the option
    >>> with dask.config.set(**{'array.slicing.split_large_chunks': True}):
    ...     array[indexer]
  return self.array[key]

Normally I would just do x.data = x.data[:, ::-1] but I'm also using rioxarray so I would like to stick to the DataArray APIs. That way rioxarray can pick up on the change.

Is there a better way to do this or at least a way to avoid oversized chunks?

[keewis]

you can invert without reindexing:

x.isel(y=slice(None, None, -1))

[fbunt]

That does the trick. Thank you! I didn't realize isel could do that.

[oscarbrousse]

When I do this rioxarray.transform() function still returns a negative increment on my Y-axis. Any solution to that?

[dcherian]

Please ask over at the rioxarray tracker.

[fbunt]

@oscarbrousse maybe try transform(recalc=True)?