Re: Making sense of 4-byte binary fields

[Scott Klement <rpg400-l@xxxxxxxxxxxxxxxx>]

> I may be going against the grain here, but if the documentation says
> 4-byte binary I just specify 4-byte binary in my code, as shown below,
> which works just fine, and the way we always used to define binary fields
> prior to RPGIV.

The problem with this is that the "B" data type is not a true binary 
field. It's a decimal field (like packed or zoned) that's stored in memory 
according to it's binary value.

Whenever you do a math operation on a "B" field, it gets converted to 
packed before the operation occurs.

For example, you provided this definition:

> 0085.00 D                 DS
> 0086.00 D  MSGDLN                 1      4B 0

That will give you a 9-digit binary field.  When any math operation is 
performed on it, it'll copy that 9-digit field to a 9-digit packed field, 
the math operation is performed on that, and then it's copied back.

The problem with that is that you can't use all 32 bits of the binary 
number. A true 4-byte (32-bit) binary number can store a value from 
-2147483648 to 2147483647. Note that that's a TEN digit number, and won't 
fit in a 9-digit packed field!

So there's two excellent reasons to use data type "I" (binary integer) 
rather than data type "B" (binary decimal)

a) Binary integer performs better.

b) Binary integer is a true BINARY(4) and supports all possible values 
that APIs can return.

The ONLY thing you have to do differntly in your code -- the absolute ONLY 
difference necessary from the programmer's perspective to gain these two 
benefits is to code the letter "I" instead of the letter "B".

For example:

     D                 DS
     D  MSGDLN                 1      4I 0


> The problem lies if you code the field using the new sexy type 'S' for a
> standalone field, when you only get the opportunity to specify the number
> of DIGITS, rather than bytes.

It's not new or sexy, it's just you define variables in RPG IV.  You 
defined variables using the number of digits in RPG II as well.

     C                     Z-ADD*ZEROS    MYNUM   70

Is that a 7 byte field? No. It's 7 digits. It's 4 bytes!  Maybe that way 
is also new and sexy! :)

But, by all means, if you find it easier to define binary fields in a data 
structure, go ahead. But PLEASE PLEASE PLEASE don't use the "B" data 
type.