fast_Herbie_xarray() does not work with hrrr's subh product about herbie HOT 3 CLOSED

awesome, thanks so much Brian! I'm a heavy user of this and find it super useful. You've done great work!

from herbie.

You're right that it is caused by reading the data into xarray. I don't know why cfgrib reads this into two separate DataFrames. That would be a good question for the cfgrib repo.

On the Herbie side, it may be possible to work around this; it would require doing something like xarray.concat(ds, dim='step') for the objects returned by hrrr subh.

from herbie.

In that last commit, Herbie will try xr.concat(xx, dim='step') on the list of Datasets returned by cfgrib. This works for the HRRR subh fields because the hypercubes have the same variable and level.

from herbie import fast_Herbie_xarray
import pandas as pd

dates = pd.date_range(
    '2022-01-01 1:00', 
    '2022-01-01 3:00',
    freq='1H'
)
fxx = [0,1,2,3]

ds = fast_Herbie_xarray(
    DATES=dates, 
    fxx=fxx, 
    model='hrrr', 
    product='subh', 
    searchString=':PRES:surface'
)
## ^ Took ~20 seconds to load

ds.isel(time=0, x=500, y=500).sp.plot(x='valid_time', marker='.')
ds.isel(time=1, x=500, y=500).sp.plot(x='valid_time', marker='.')
ds.isel(time=2, x=500, y=500).sp.plot(x='valid_time', marker='.')

@isodrosotherm Thanks for reporting this. These comments always help me improve Herbie.

from herbie.