NOTICE: This version of the NSF Unidata web site (archive.unidata.ucar.edu) is no longer being updated.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.
NOTE: The netcdf-hdf
mailing list is no longer active. The list archives are made available for historical reasons.
Chris, > 1) The netCDF manual doesn't explicitly forbid it, but from the > code it looks impossible for a string valued attribute > to have more than one value, i.e. in CDL: > > my_var:greetings = "hello", "goodbye"; > > Am I correct that this is not allowed? You can do this in CDL, but it just concatenates the strings into a single array of characters. Since any attribute is just a one-dimensional array of values of some type, the result is an attribute of characters of length large enough to hold the string "hellogoodbye". There is really no such thing as a "string-valued" attribute, since "string" is not one of the primitive netCDF types. However multi-valued attributes of character type are not very different from string-valued attributes, in that they are an array of characters with a length. If you want to adopt a convention that "," separates strings, for example, then the attribute "one,two,three" could be thought of as a string-valued attribute containing three strings, but that's purely convention. The netCDF library treats it as a single attribute of 14 characters. > 2) We have been building off of netCDF 2.02 on your ftp > server. Have there been any bug fixes / modifications > since then that we should be aware of? Glenn described a significant optimization for netCDF on UNIX that we haven't decided how to package yet. There was a minor change to ncdump and ncgen to permit the netCDF names in a CDL file to begin with a non-alphabetic character, and thus to permit ncgen to work for files generated by ncdump from filenames that begin with a non-alphabetic character. There have been some clarifications and fixes to the documentation and the installation instructions and a minor bug fix to the FORTRAN interface. I don't think you need to be concerned with any of these changes for your release. --Russ
netcdf-hdf
archives: