NCL Home > Documentation > Functions > Type converters

short2flt

Converts values of type short to values of type float using the "scale" and "offset" attributes (if present) in accordance with the CF and COARDS netCDF conventions..

Prototype

load "$NCARG_ROOT/lib/ncarg/nclscripts/csm/contributed.ncl"

	function short2flt (
		x  : short   
	)

	return_val [dimsizes(x)] :  float

Arguments

x

An array of type short of any dimensionality.

Return value

The results are returned in a float array of the same type dimensionality as x. Metadata is preserved.

Description

This function converts short data to float using the "scale" and "offset" attributes (if present), and preserves metadata. It assumes that the data are first scaled and then the offset is applied:

             x_float = x_short*scale + offset

This is in accordance with the CF and COARDS netCDF conventions. [See: http://www.unidata.ucar.edu/software/netcdf/conventions.html ]

Other attributes similar to "scale" and "offset" will be recognized:

offset:

"add_offset", "OFFSET", "Offset", "_offset", "Intercept", "intercept"

scale:

"SCALE", "Scale", "_scale", "scale_factor", "Scale_factor", "Slope" , "slope"

Note1: This function is slightly different than the built-in set of conversion functions (like floattointeger and shorttobyte), because it requires loading the "contributed.ncl" script and it preserves metadata. It has a slightly different naming style than the built-in functions, just to set it apart from these other functions.

Note2: A separate function, short2flt_hdf, must be used if the the following approach is to be applied:

             x_float = x_short*(scale - offset)

See Also

short2flt_hdf, dble2flt, flt2dble, byte2flt, typeof, getfilevartypes

Examples

Example 1

The file "air.2m.gauss.1979.nc" contains temperatures at 2 meters. The data are store as type "short" with a scale_factor and add_offset:

        short air(time, level, lat, lon) ;
                air:long_name = "Daily Forecast of Air temperature at 2 m" ;
                air:units = "degK" ;
                air:add_offset = 447.65f ;
                air:scale_factor = 0.01f ;
                [SNIP]

Convert to type "float" as the data are being read:

 load "$NCARG_ROOT/lib/ncarg/nclscripts/csm/contributed.ncl" 

begin
  f    = addfile("air.2m.gauss.1979.nc" , "r")
  t2m  = short2flt( f->air )  
  printVarSummary(t2m)
end

The output contains

Variable: t2m
Type: float
Total Size: 26350080 bytes
            6587520 values
Number of Dimensions: 4
Dimensions and sizes:   [time | 365] x [level | 1] x [lat | 94] x [lon | 192]
Coordinates: 
            time: [1569072..1577808]
            level: [ 2.. 2]
            lat: [88.542..-88.542]
            lon: [ 0..358.125]
Number Of Attributes: 18
  long_name :   Daily Forecast of Air temperature at 2 m
            [SNIP]

NOTE 1: it is possible to test a variable's type prior to reading the variable to memory. Using the above example:

  f    = addfile("air.2m.gauss.1979.nc" , "r")
  if (getfilevartypes (f,"air").eq."short") then)
      t2m  = short2flt( f->air )
  else
      t2m  = f->air 
  end if

NOTE 2: The above file's time units are "hours since 1800-1-1 00:00:0.0". Many people find these units rather cumbersome to use. To convert to more 'human readable' time units the ut_calendar function can be used:

   ymdh = ut_calendar(time, 0)          ; ymdh(:,0:5)
   yyyymmddhh = ut_calendar(time, -3)   ; integer