The ABL DECIMAL is the ABL data type that can represent the widest range of .NET System.Double and System.Single values, and the reverse is also true. However, the ABL DECIMAL cannot represent all values of the .NET System.Double, and the .NET System.Single cannot represent all values of the ABL DECIMAL, as the examples in this topic show.

The following example raises a run-time error because the maximum System.Double value represents a whole number of over 300 digits that is too large to assign to an ABL DECIMAL that can hold a maximum of 50 digits:

DEFINE VARIABLE dTest AS DECIMAL NO-UNDO.
dTest = System.Double:MaxValue. /* Run-time error */

The following example raises a run-time error because the negative value of the ABL DECIMAL (50 whole-number digits), is too large to pass to a System.Single method parameter, which can represent whole-number values of no more than 38 digits:

DEFINE VARIABLE dResult AS DECIMAL NO-UNDO.
DEFINE VARIABLE dTest   AS DECIMAL NO-UNDO.
ASSIGN
  dResult = -99999999999999999999999999999999999999999999999999
  dTest   = System.Math:Abs(dResult AS FLOAT).
 /* Run-time error */

The following example returns a value for dResult of 0.0000003469 (with rounding), because an ABL DECIMAL can only represent 10 digits to the right of the decimal point. Any remaining digits are truncated:

DEFINE VARIABLE dResult AS DECIMAL NO-UNDO.
dResult = System.Math:Abs
          ((34687984589034945 / System.Math:Pow(10, 23)) AS DOUBLE).

Note that If the ABL calculation and DECIMAL data type maintained the full precision of the .NET System.Double, the value of dResult would be 0.00000034687985 (with rounding).