Reputation: 81
I want to select a Dataarray according to a criteria from a non-dimension coordinate. In the example below, I have my coords 'sat' which depend on the coord 'time', which is exactly what I need. When I use Dataarray.sel(time='2021-05-04')
, everything works fine as expected, but I need to be able to select with this criteria Dataarray.sel(sat='L30')
. I could be open to stack the info about 'sat' an other way if it's more convenient for the query, but I don't want to separate my dataarray in two dataarray (sat='L30 and 'S30') merged in a dataset.
Thanks!
<xarray.DataArray (band: 5, time: 48, lat: 93, lon: 82)>
array([...])
Coordinates:
* band (band) <U5 'Fmask' 'blue' 'nir_n' 'red' 'swir2'
* lat (lat) float64 5.415e+06 5.415e+06 5.415e+06 ... 5.413e+06 5.413e+06
* lon (lon) float64 6.658e+05 6.658e+05 6.659e+05 ... 6.682e+05 6.682e+05
* time (time) datetime64[ns] 2021-05-04T15:35:59 ... 2021-10-28T15:26:29
sat (time) <U3 'S30' 'L30' 'S30' 'S30' ... 'S30' 'S30' 'S30' 'S30'
Attributes:
transform: (30.0, 0.0, 665790.0, 0.0, -30.0, 5415330.0)
crs: +init=epsg:32619
res: (30.0, 30.0)
is_tiled: 1
nodatavals: (-9999.0,)
scales: (1.0,)
offsets: (0.0,)
AREA_OR_POINT: Area
Upvotes: 3
Views: 1604
Reputation: 81
Finally, found the elegant enough answer in an old issue on xarray github:
ds.sel(time = ds.sat == 'L30')
Notes:
ds = the DataArray (or Dataset)
time : because my sat coordinate is linked to the time coordinate
ds.sat exists because I created it
Upvotes: 5