I have a work divided in two tasks : one program is sorting model data (ERA5 from ECMWF) and satellite observations and write the selected one in hdf5 files, the second program is using these files to perform simulations.
In the begening I was writing the files from the first task in netcdf v3 but It start to be diffecult to read them with some software. in netcdf4 I have a conflict of libraries in the header since that I cannot identify and fix since I do not have the root user on the server me are using.
So I change to write the files in pure hdf5, that actually find simpler than netcd4. However, it is not working.
It works If I do an “empty” run ; meaning I create the tables to save fill them with zeros instead of the model values. But when I do a real run I try to open with panoply I get the error “There was an error opening the dataset: DataObject doesnt start with OHDR”
I am guessing maybe there is bad values in the tables, but not sure because I cannot access to the data. Is there an option that could help ?
Thanks for the help.