This archive contains answers to questions sent to Unidata support through mid-2025. Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated. For the most up-to-date information on the use of NSF Unidata software and data services, please consult the Software Documentation first.
Hi Hannes, > I create a nc file with dimensions x = 162 and y = 76 > with a daily timestep for 10 years, 4 variables ( 1 real, 3 integer) > How large should such a file be.... > because just by creating it grows up to 568 MB. > Is such size reasonable ? seems to be to large to me.. Each real or integer variable value takes 4 bytes. So just the size to store the values is about 162*76*360*10*4*4 or about 709 MB. If you are using the unlimited dimension for time, then the file size should only reflect the number of time steps you have actually written. So after writing one time step, the storage space for values in the file should only be about 162*76*4*4 or about 200 KB. For questions like this, it is helpful to send the structure of the netCDF file, which you can get by running ncdump -h FILENAME > Second question > it is possible to open an netcdf file and write from several processes to it.. > I know already that several processes can have read access to one file, > with one process writing... With netCDF-3, there should only be one writer. If you need to have multiple processes writing the same file concurrently, you can use Parallel netCDF: http://www-unix.mcs.anl.gov/parallel-netcdf/ or wait until the release of netCDF 4, which will also support parallel writing to a single file through MPI-IO in HDF5. --Russ _____________________________________________________________________ Russ Rew UCAR Unidata Program address@hidden http://www.unidata.ucar.edu