functions in netcdf.i - n

             nc_addrec, f, time  
 
       -or- nc_addrec, f  
     adds a new record to the netCDF file F at time TIME.  
interpreted function, defined at i/netcdf.i   line 501  

             nc_attrdef, ncf, attr_name, var_name, value  
 
     sets the value of the netCDF attribute ATTR_NAME associated  
     with variable VAR_NAME to VALUE (note that the data type of VALUE  
     becomes the data type of the attribute).  
     The NCF is the structure returned by nc_create; nc_attrdef  
     must be called prior to nc_enddef, which actually writes the  
     attribute data to the file.  
     If VAR_NAME is omitted, ATTR_NAME refers to the whole file.  
interpreted function, defined at i/netcdf.i   line 395  

             value= nc_attribute(attr_name, var_name)  
 
     gets the value of the netCDF attribute ATTR_NAME associated  
     with variable VAR_NAME, or nil if none.  Uses the external  
     variable nc_file set by nc_open.  
     If VAR_NAME is omitted, ATTR_NAME refers to the whole file,  
     and is retrieved (if present) from the nc_file.attrs variable.  
interpreted function, defined at i/netcdf.i   line 514  

             ncf= nc_create(filename)  
 
     creates a netCDF file FILENAME.  
     After this call, use nc_vardef to declare the netCDF variables.  
     Then use nc_enddef to write the netCDF self-descriptive  
     information.  Only after this are you free to actually write data.  
interpreted function, defined at i/netcdf.i   line 238  

             nc_dimdef, ncf, dim_name, size  
 
       -or- nc_dimdef, ncf, dim_name, "unlimited"  
     define a named dimension.  The SIZE parameter is the length of  
     the dimension, or the string "unlimited" for the unlimited  
     dimension.  (The numerical value 0 is the same as "unlimited".)  
     You can also define named dimensions implicitly using nc_vardef.  
interpreted function, defined at i/netcdf.i   line 368  

             def_string= nc_dimsof(var_name)  
 
     returns the dimension list of a netCDF variable VAR_NAME in symbolic  
     form, i.e.- using the netCDF dimension names.  This requires the  
     nc_file external variable set by nc_open.  
interpreted function, defined at i/netcdf.i   line 539  

             f= nc_enddef(ncf)  
 
     creates netCDF file NCF (returned by nc_create), and writes the self-  
     descriptive information.  Returns the ordinary Yorick file object  
     corresponding to the new file.  You are then free to write variables,  
     or use the save or nc_addrec functions.  
interpreted function, defined at i/netcdf.i   line 424  

             f= nc_open(filename, mode)  
 
     opens a netCDF file FILENAME for reading or update as specified  
     by MODE, which defaults to "rb".  Attributes and dimension names  
     can be found in the three external variables nc_dims (an array of  
     type NC_dim), nc_attrs (an array of type NC_attr), and nc_vars  
     (an array of type NC_var) after this call.  
     MODE should be either "rb" or "r+b"; nothing else makes sense.  
     If FILENAME is an array of strings, exactly those files will be  
     opened as a family (if possible).  Note that nc_open("myfile00")  
     potentially opens myfile01, myfile02, and so on, as for openb,  
     but that nc_open(["myfile00"]) opens myfile00 only.  
interpreted function, defined at i/netcdf.i   line 36  

             nc_vardef, ncf, name, type, dims, record=0/1  
 
       -or- nc_vardef, ncf, name, type, record=0/1  
       -or- nc_vardef, ncf, name, template=template, record=0/1  
     define a variable in the NCF (returned by nc_create) with name  
     NAME, type TYPE (as returned by typeof or structof), and dimensions  
     DIMS (as returned by dimsof).  The template= keyword may be used  
     instead of type and dims; the type and dims will be those of the  
     TEMPLATE.  If dims is not specified, a scalar is assumed.  If the  
     record= keyword is present and non-zero, the variable is a record  
     variable; otherwise it is a non-record variable.  
     You can use the dimnames= keyword to write specific dimension  
     names into the netCDF file.  These are not useful to Yorick, but  
     other codes may require them.  If two variables share a dimension  
     name, the corresponding dimension must have the same length.  For  
     example:  
       nc_vardef, ncf, "theta", double, [1,nlat], dimnames=["latitude"]  
       nc_vardef, ncf, "phi", double, [1,nlong], dimnames=["longitude"]  
       nc_vardef, ncf, "elevation", double,  
                  dimnames=["latitude","longitude"]  
     A dimension name of "" lets Yorick invent a fake dimension name,  
     as it does by default.  If dimnames= is present and the lengths  
     of the dimensions have previously been defined, then the DIMS  
     parameter is unnecessary, as in the "elevation" array in the example.  
     You can use the nc_dimdef function to define a named dimension size  
     before you define any variables with that dimension.  
interpreted function, defined at i/netcdf.i   line 252  

             nc_open, nc_create, nc_vardef, nc_enddef, nc_addrec  
 
     are the main routines to read and write netCDF files.  
     The ordinary openb function will also open netCDF files.  
     Writing a netCDF file is more problematic in Yorick, since  
     you must define the entire file structure before you write  
     any data.  Therefore, the nc_create call returns only a  
     "token" for nc_vardef, which you use to declare variables  
     in the file.  When you are done declaring variables, you  
     call nc_enddef, which returns an ordinary Yorick file object.  
     You can then write data to the file (with f.var=value or  
     save,f,var).  To add a record, you must use nc_addrec instead  
     of add_record (nc_addrec updates the record count in the file).  
keyword,  defined at i/netcdf.i   line 17