You can use a literal as a value expression. A literal is either
    a character string or a numeric literal.

    Numeric literals can take the following forms:

    o  A decimal string consisting of digits and an optional decimal
       point. The maximum length, not counting the decimal point, is
       19 digits.

    o  A decimal number in scientific notation (E-format), consisting
       of a decimal string mantissa and a signed integer exponent,
       separated by the letter D (for double), E (for E-format) or Q
       (for H_floating).

    CDO allows you to use unary plus and minus signs in numeric
    literals. Numeric literals must start and end with a numeral and
    cannot include hexadecimal digits. Numeric literals in E notation
    cannot include embedded spaces.

    The following expressions are valid numeric literals:

       +123
       -3.49
       0.3338889909
       6.03 E+23

    If you use a numeric literal to assign a value to a field or a
    variable, the data types of the field or variable determine the
    maximum value you can assign.

    A period at the end of a data definition command line terminates
    the command; therefore, you cannot use a decimal point to
    terminate a number if you want to include more data definition
    clauses in the statement.

    If you want to include more data definition clauses, include a
    zero after the decimal point, or place the value expression in
    parentheses:

       COMPUTED BY X * 2.0
       COMPUTED BY (X * 2.)