This example uses Haversine's formula to find the great circle distance between two points assuming a mean earth radius of 6371 kilometers. The example is published in Wikipedia. 

1. The latitude and longitude for Los Angeles International Airport (LAX) is N 33.94°, W 118.4°. In radians, these values are .5924 and -2.0665. The latitude and longitude for Nashville International Airport (BNA) in Nashville, Tennessee is N 36.12°, W 86.67°. In radians, these values are .6304 and -1.5127.  Create an input array to contain these values:

   AFL% store(build(<airport:string, lat:double, lon:double>[i=0:1],
                    '[(\'LAX\', .5924, -2.0665), (\'BNA\', 0.6304, -1.5127)]', true), airports);

   The output is:

   {i} airport,lat,lon
   {0} 'LAX',0.5924,-2.0665
   {1} 'BNA',0.6304,-1.5127

2. The input attribute arguments for the gcd_hf function must be from the same cell.  From the airports array, use filter and cross_join to build a trip array describing the airport entries we care about.  The cast operator is used to rename the lat and lon attributes:

   AFL% store(cast(cross_join(filter(airports, airport='LAX'),
                              filter(airports, airport='BNA')),
                   <from:string, latf:double, lonf:double, to:string, latt:double, lont:double>[i;j]),
              trip);

   The output is:

   {i,j} from,latf,lonf,to,latt,lont
   {0,1} 'LAX',0.5924,-2.0665,'BNA',0.6304,-1.5127

3. Use apply and gcd_hf to add the Haversine great circle distance attribute:

   AFL% apply(trip, distance, gcd_hf(6371.0, latf, lonf, latt, lont));

   The output is:

   {i,j} from,latf,lonf,to,latt,lont,distance
   {0,1} 'LAX',0.5924,-2.0665,'BNA',0.6304,-1.5127,2886.43

4. Remove the example arrays:

   AFL% remove(airports); remove(trip);