Re: SFLRCDNBR

On 04/06/2007, at 6:58 AM, SirJohn wrote:

You know with file sizes reaching well over 100M and you can have 1TB on the
box, you would think that to display more that 9,999 records and reference
them would be no big deal.

Because a sub-file that big indicates a design flaw in the application building it. Further, if you really need a sub-file that large then an acceptable alternative exists in UIM. A UIM list can grow to 16MB of data (and is much more flexible than a sub-file)

And what about the source limit? I was working
on a JDE program and tried to add a line and guess what? You can have like
32, 726 lines of code.

Again, a program with that much source in a single member is poorly designed. It is crying out to be modularised.

The 32KB limit is not a system restriction. A source member can have millions of records. The limit is imposed by SEU and IBM's view of that is you should not be using that for modern development. You should be using WDSc or another client-based development environment.

Why does this box limit everything to that 32K?

It doesn't. Most things are limited to 16MB as a side-effect of the addressing scheme. However, even that limit is not absolute. 16MB chunks can be strung together to appear larger. Teraspace is available if you really need a contiguous 4GB chunk. Refer to the published database limits and you'll see this system supports much larger limits than almost any other commercial system.



Regards,
Simon Coulter.
--------------------------------------------------------------------
FlyByNight Software OS/400, i5/OS Technical Specialists

http://www.flybynight.com.au/
Phone: +61 3 9899 0985 Mobile: +61 0411 091 400 /"\
Fax: +61 3 9899 0985 \ /
X
ASCII Ribbon campaign against HTML E-Mail / \
--------------------------------------------------------------------