From: Howard Brazee on
On Fri, 11 Jan 2008 20:01:56 GMT, "William M. Klein"
<wmklein(a)nospam.netcom.com> wrote:

> Aren't you using IBM COBOL? If so, you can (and I know I have posted this
>before). It isn't FASTSRT - or even totally "obvious" - but it does use DFSort
>or SyncSort to do the sorting. See:
>
>http://publibz.boulder.ibm.com/cgi-bin/bookmgr_OS390/BOOKS/igy3pg32/1.12.10.2
>
>Yes, I would like IBM to add the "native" table sort feature, but what you ask
>for is possible.
>
>P.S. At least on mainframes (IBM or otherwise) I would expect this technique
>often to use the system's "optimal" sort technique for the data involved.

Our compiler options are set the same for all programs when they get
migrated to production.
From: Frank Swarbrick on
>>> On 1/11/2008 at 1:28 PM, in message
<09kfo39k2ed2uoag56np597v4mj6g4ao38(a)4ax.com>, Howard
Brazee<howard(a)brazee.net> wrote:
> On Fri, 11 Jan 2008 20:01:56 GMT, "William M. Klein"
> <wmklein(a)nospam.netcom.com> wrote:
>
>> Aren't you using IBM COBOL? If so, you can (and I know I have posted
> this
>>before). It isn't FASTSRT - or even totally "obvious" - but it does use
> DFSort
>>or SyncSort to do the sorting. See:
>>
>>http://publibz.boulder.ibm.com/cgi-bin/bookmgr_OS390/BOOKS/igy3pg32/1.12.1

> 0.2
>>
>>Yes, I would like IBM to add the "native" table sort feature, but what
> you ask
>>for is possible.
>>
>>P.S. At least on mainframes (IBM or otherwise) I would expect this
> technique
>>often to use the system's "optimal" sort technique for the data involved.
>
> Our compiler options are set the same for all programs when they get
> migrated to production.

What happens if you put a PROCESS or CBL card at the top of your source
file? Will production implementation reject it because you are trying to
use a non-standard compile option?


From: William M. Klein on
Huh???
Compiler options have nothing to do with this. For IBM mainframes, the code
from the programming guide will work the same (and use the "system sort program"
to sort tables) regardless of compiler option.

--
Bill Klein
wmklein <at> ix.netcom.com
"Howard Brazee" <howard(a)brazee.net> wrote in message
news:09kfo39k2ed2uoag56np597v4mj6g4ao38(a)4ax.com...
> On Fri, 11 Jan 2008 20:01:56 GMT, "William M. Klein"
> <wmklein(a)nospam.netcom.com> wrote:
>
>> Aren't you using IBM COBOL? If so, you can (and I know I have posted this
>>before). It isn't FASTSRT - or even totally "obvious" - but it does use
>>DFSort
>>or SyncSort to do the sorting. See:
>>
>>http://publibz.boulder.ibm.com/cgi-bin/bookmgr_OS390/BOOKS/igy3pg32/1.12.10.2
>>
>>Yes, I would like IBM to add the "native" table sort feature, but what you ask
>>for is possible.
>>
>>P.S. At least on mainframes (IBM or otherwise) I would expect this technique
>>often to use the system's "optimal" sort technique for the data involved.
>
> Our compiler options are set the same for all programs when they get
> migrated to production.


From: Robert on
On Fri, 11 Jan 2008 08:16:44 -0700, Howard Brazee <howard(a)brazee.net> wrote:

>On Fri, 11 Jan 2008 04:51:38 GMT, Arnold Trembley
><arnold.trembley(a)worldnet.att.net> wrote:
>
>>Here's an interesting article on table sorting in COBOL. The author
>>includes an example of Quicksort written in COBOL using iteration
>>rather than recursion:
>>http://home.att.net/~arnold.trembley/svalgard.htm
>
>I have programmed the Quicksort in CoBOL to sort some large tables.
>It's fast, easy, and clear. That said, there should be no need for
>it - I should be able to CoBOL to sort it, and CoBOL should be able to
>tell the system to pick the best sort.
>
>But my shop's CoBOL can't do that.

Doesn't your shop's CoBOL support object oriented?
From: Pete Dashwood on


"Robert" <no(a)e.mail> wrote in message
news:r7kdo3pfq09catidnsddj0favs4e0uqqsj(a)4ax.com...
> On Thu, 10 Jan 2008 08:20:35 -0600, "HeyBub" <heybub(a)gmail.com> wrote:
>
>>Robert wrote:
>>> But then, if it follows the 2002 Standard
>>> for SORT verb, you can sort a table in memory with one line of code.
>
>>Or use the SORT verb and let the operating system figure it out.
>
> You must be referring to the SyncSort operating system. I booted it on my
> Genius (as seen
> on TV). Now it not only chops, slices and dices, but also sorts the pieces
> into neat
> little heaps. That was on the Heapsort setting, of course.
>
> Then I took it to the laundry room, set it on Bubblesort, and dumped my
> clothes in. It
> worked, but performance sucked. It ran in n^2. The book said I could
> expect n(log n).
> Since there were only two buckets, white and colored, would Bucketsort
> have run faster?
>
> I couldn't find Coinsort, so I tried sorting my change with Shell Sort. It
> threw them all
> into the reject pile. Next time I'll try Pigeonhole or Bozo Sort.

ROFL!

Great job, Robert!

Thanks, I needed that... :-)

Pete.
--
"I used to write COBOL...now I can do anything."