读取grb2文件很慢

Question

Reading a grb2 file of about 100 megabytes takes me 5 minutes with the following code:

import cfgrib
datasets = cfgrib.open_datasets(grib_path, backend_kwargs={'indexpath': ''})

Is there any way or library with which I can read these files more quickly?

Answer 1

You can activate parallel read to enhance the reading speed:

import xarray
data = xarray.open_dataset(grib_path, engine='cfgrib', parallel=True)

Answer 2

it's actually an annoying issue that's being raised here. With a 12Mb Grib file, to read a single field, cfgrib takes 19 seconds compared to 0.02 seconds with pynio! more than 900 times more!

ps : parallel = true gives as result : open_dataset() got an unexpected keyword argument 'parallel'