NOTICE: This version of the NSF Unidata web site (archive.unidata.ucar.edu) is no longer being updated.
Current content can be found at unidata.ucar.edu.
To learn about what's going on, see About the Archive Site.
Hi netCDF group, I was wondering if the developers would consider enhancing ncgen to allow for binary data? Currently, ncgen accepts only ASCII data: data: VAR = 0.5, 1.5, 2.5, 3.5, 4.5, 5.5, 6.5, . . . I thought it would be nice if the format allowed for an external binary file in such a way as data: VAR: file("path/to/mydata.bin", stream, big_endian) [I don't claim this is the best syntax. This is just for illustration of the idea.] I'm trying to find a tool with which it's easy to convert a stream-binary file into a netCDF format. Then, it struck me: Why, ncgen must be the best tool! So for, Ferret comes close, but it doesn't allow for large binary files and the Ferret language takes a bit getting used to. I'm aware of tools that can convert a pair of a GrADS control file and a stream binary into a netCDF file, but they can handle only single-precision floats. I know that you can write a C or Fortran program to do that, but using the C and Fortran API is too much work (dim-IDs, var-IDs, put_att, enddef, NF_FLOAT, etc. etc.), compared to the simplicity of the CDL format or the GrADS control file format. <netcdfgroup@xxxxxxxxxxxxxxxx>I think this is the very reason very few scientists around me write programs to generate netCDF files. Most of them simply use GrADS with stream binary files because it's so easy to write GrADS control files. If you can use the CDL+ncgen combination to generate a netCDF file, I'd be able to convince them to switch to the netCDF format. Best regards, Ryo
netcdfgroup
archives: