The current implementation of xarray loads the whole dataset in memory, which can be very large depending on the gridding, area and time steps.
a lazy-loading implementation (open_dataset) or an implementation modifying the file in place (netCDF4) would be preferable.
The current implementation of xarray loads the whole dataset in memory, which can be very large depending on the gridding, area and time steps.
a lazy-loading implementation (open_dataset) or an implementation modifying the file in place (netCDF4) would be preferable.