Re: int vs binary questions

[Hans Boldt <boldt@xxxxxxxxxx>]

James Rich wrote:
> On Mon, 9 Feb 2004, Hans Boldt wrote:
>
>
> > Note that any operations with binary decimal numeric type involve an
> > implicit conversion to packed decimal numeric type. That will slow down
> > any operation involving binary.
>
>
> Is this only with binary type, or does this also happen with integer type?
> And, what is the difference between integer and binary?  AFAIK, both int
> and binary store numbers as a straightforward binary representation, i.e.
> 0001 for 1 or 0101 for 5 (of course with leading zeros to fill up 4
> bytes).

The implicit conversion to packed decimal happens only with binary 
decimal format (and zoned decimal format). Operations on integer and 
unsigned values occur in integer or unsigned form whenever possible. 
(However, if an operand of a numeric expression is decimal, the operands 
are converted to decimal. Likewise, if an operand of a numeric 
expression is in float format, the other operands are converted to 
float. Check the Reference manual for further details.)

Yes, both binary decimal and integer represent numbers in two's 
complement form. But binary decimal caps the values to a range of 
-999999999 to 999999999 to assure functional compatibility with the 
other decimal types. Integer format allows the full range of two's 
complement values.

Cheers! Hans