NCL Home > Documentation > Functions > File IO


Returns all dimension names associated with a variable on a supported file.


	function getfilevardims (
		thefile [1] : file,    
		varname [1] : string   

	return_val [*] :  string



A reference to a file created from a call to addfile or addfiles. Thus, the file referenced must be one in the supported file format list.


The name of the variable in single string format.

Return value

Returns a one dimensional array of dimension names for a file variable. getfilevardims returns a missing value if no dimension names exist or the variable is not defined.


Accesses dimension names of a user specified variable in a supported file format.


Consider a netCDF file where "ncdump -h" yields:

netcdf Tstorm {
        timestep = 64 ;
        lat = 33 ;
        lon = 36 ;
        timelen = 20 ;
        float t(timestep, lat, lon) ;
                t:_FillValue = -9999.f ;
        int timestep(timestep) ;
        float lat(lat) ;
        float lon(lon) ;
        char reftime(timelen) ;
                reftime:units = "text_time" ;
                reftime:long_name = "reference time" ;
To retrieve the dimension names of the variable "t" on the file, you can use the following code snippet:
   f     = addfile ("$NCARG_ROOT/lib/ncarg/data/cdf/Tstorm.cdf" , "r")
   dNames= getfilevardims(f,"t")  
   print (dNames)
This produces the output:
Variable: dNames
Type: string
Total Size: 12 bytes
            3 values
Number of Dimensions: 1
Dimensions and sizes:   [3]
(0)     timestep
(1)     lat
(2)     lon
Note: you can also accomplish the above with the following code snippet:
   f     = addfile ("$NCARG_ROOT/lib/ncarg/data/cdf/Tstorm.cdf" , "r")
   dNames= getvardims(f->t)  
   print (dNames)
This is more inefficient, however, because the reference "f->t" causes the variable "t" to be read off the file. In using "getfilevardims", only the dimension names are read off the file.