Hi Ian,

I was thinking more of the subroutine call versus the 002: COUNT(... approach. I've always found subroutine calls slower, at least in 3.x.x. I've never had good luck with the A type 008 CALL.

I'm not looking to upgrade anytime soon. I suppose I'll have to eventually, but I'm putting it off as long as I can, and when I do, it will likely be to 4.x.

Charlie

On 07-21-2011 3:34 AM, inavran wrote:
Hi Charlie,

I haven't done a performance test, but I've been informed by Dan Klein
(who has helped me alot in the past, and i trust) that calling a
global subroutine within an I-Type dictionary will perform much better
than using a A-type; 008: CALL mysub or 008:B;mysub (which ive found
very slow in the past, and therefore wont use), i sure Jim will say if
im right or wrong?

Those PDFs i found on the web has been very informative reading. Its
not specific to jBASE, but alot of the commands work, and it filled in
alot of gaps to do with A-type DICTs

Here's a link to a public goggle docs share : http://tinyurl.com/3vly68j

I dont think i will be putting the time in to test it out, as we will
be migrating to jb 5.2.x in the coming months (ive been saying that
for 6 months already though!). jb3 has served us well, but its time to
move on.

Do you have any plans to migrate soon?

Ian




On Jul 20, 7:32 pm, Charlie Noah <[email protected]> wrote:
Hi Ian,
I'd love to get a copy of that PDF. Let us know where you end up putting it so I can grab a copy.
Have you done a performance test to determine if the subroutine is faster or slower than the I-descriptor with COUNT? I'd be very interested in your results. I'm glad to know I'm not the only one still running 3.x.
Thanks,
Charlie
On 07-20-2011 11:57 AM, inavran wrote:Thanks everyone for your comments. I did put this to devsup as well and with help from Dan Ell & Dan Klein, a more elegant solutions was found. Thanks for both of them as well. see below. Dan Ell's First offered this solution using an I descriptor 001 I 002 DCOUNT(@RECORD<1>,’,’) 003 004 Commas in 1 005 8R 001 I 002 DCOUNT(@RECORD<3>,@VM) 003 004 VM in 3 005 8R This didn't work for me first time (as DCOUNT subroutine doesn't exist on jBASE 3.x, Dan thought i was running jb4/5) Dan Klein, then kindly suggested all i needed to do was to create the DCOUNT sub, and catalog it so it would be available system-wide 001 SUBROUTINE DCOUNT(result, param1, param2) 002 result = DCOUNT(param1, param2) 003 RETURN This did the trick nicely.... Over to Jim's point. Im sorry, but with respect, i think you've got completely wrong end of the stick. If i was storing information for db purposes it would be a terrible model, the fact is, i was
 storing it as a quick dirty report i needed (that i was too lazy to write another program for, when i just wanted a quick ref count on how many MVs i had from each key. (ie the majority would be a count of 1 and a small few would be > 1), i would simply then select the keys > 1 and see what sort of data set i was dealing with. I put the commas count eg in as literally an example (be it a very bad example in hindsight) - nothing else. I would never store multiple values separated by commas in a data model. (even if i did, i wouldn't be silly to mention it on this a board)... Thanks for your reply though, a always enjoy reading your posts :) Charlie, i found an excellent 4 page pdf (scrapped off the spectrum website years ago), which describes in detail all the DICTs & ACCESS commands (similar to JQL). It lists all TCL commands and uses of correlatives (for A types not I) If anyone would like a copy of it, let me know and i can put it on a public dropbox link. Ian

    

--
Please read the posting guidelines at: http://groups.google.com/group/jBASE/web/Posting%20Guidelines
 
IMPORTANT: Type T24: at the start of the subject line for questions specific to Globus/T24
 
To post, send email to [email protected]
To unsubscribe, send email to [email protected]
For more options, visit this group at http://groups.google.com/group/jBASE?hl=en

Reply via email to