float_type ::= "number" | "float" | "real" | ( "double" "precision" )
This rule is called by datatype
datatype