Due to the lack of model data that includes NaNs this use case is not tested so far.
Because model data witrh NaNs is quite common (regional model, satellite data, model experiments starting not on year boundaries) this is actually more common than it sounds.
a possible small dataset is /lustre/storeA/project/aerocom/aerocom-users-database/C3S-Aerosol/SLSTR_SU_v1.12/renamed/aerocom3_SLSTR_SU_v1.12_od550aer_Surface_2017_daily.nc (although the naming is slighly wrong
Due to the lack of model data that includes NaNs this use case is not tested so far.
Because model data witrh NaNs is quite common (regional model, satellite data, model experiments starting not on year boundaries) this is actually more common than it sounds.
a possible small dataset is
/lustre/storeA/project/aerocom/aerocom-users-database/C3S-Aerosol/SLSTR_SU_v1.12/renamed/aerocom3_SLSTR_SU_v1.12_od550aer_Surface_2017_daily.nc
(although the naming is slighly wrong