approximate_num_type

approximate_num_type 
      ::= 
      "float" [ "(" precision ")" ] 
      | "real" 
      | "double" "precision" 

This rule is called by

-> num_type