QUICK procedure problem.

[pickerij@norbord.com]

Yikes!! (technical term expressing shock and dismay :-)

The thought of reading sequentially through 100,000 records with Quick   
makes me gasp. Sounds like something one of those big expensive   
proprietary RDBM's would do behind your back :-)

But then I look at your code and you seem to be doing the "while   
retrieving" using a key (MYKEY). The next question is "how many records   
have the same value for THISKEY?" If the answer is "lots" (as defined by   
your patience and machine size :-) then this task really should be taken   
off line and given to Qtp.

The next problem is your apparent loss of chain info. Just how is Quick   
supposed to know what to do with MYFILEALIASED? You should read the   
record in this file via a *unique* key using the value from the designer   
file. And the designer file really should have a select to discard   
records for which the value of "NONKEYFIELD" is not "XYZ" so you don' t   
keep reading forever.

Although you indicate that you are updating NON key data, the only reason   
to use 2 files with a put and a delete is because you are actually   
updating key information or information on which keys depend such as sort   
fields or even extended sort fields -- back to the Image manual with you!   
If it was truly an update on NON key data then one file would be   
sufficient.

Sigh, so many to teach, so little time,
John Pickering
Toronto


 -----Original Message-----
From: shuckvale@cosworth-racing.co.u
Sent: Tuesday, February 23, 1999 4:37 PM
To: powerh-l@lists.swau.edu
Subject: QUICK procedure problem.
 ----------------------------------
All,


Can anyone help me with a piece of QUICK screen code (current version   
7.09E
on MPEix) which is causing no end of problems to me.  Currently I have
several screens running a while retrieving sequential construct to   
perform
updates on selected records from an IMAGE dataset with over 100,000   
records.
These screens were fine in the early days when the dataset in question   
was
smaller but now they're a major drain on CPU time which I need to plug.   
 My
current (non-)solution is the code below;


********************************************************


file MYFILE designer open 1
file MYFILE alias MYFILEALIASED open 2
...
procedure internal UPDATEMYFILE
begin
  while retrieving MYFILE viaindex MYKEY using THISKEY
  begin
    let NONKEYFIELD of MYFILEALIASED = "XYZ"
    put MYFILEALIASED reset
    delete MYFILE
    put MYFILE
  end
end


********************************************************


This seems to work fine until the 'while retrieving' loop hits the   
records
added via the alias and (I think) loops infinitely.  Without the alias I
only get to update one record, as soon as I PUT to the dataset the record
pointers get lost and the while retrieve construct falls through.


My apologies in advance if this is elementary stuff, I'm afraid it's
inherited code and whilst I've done the 'Intro to Powerhouse' course I'm
still waiting to get on 'Advanced Part 1' which I believe covers QUICK   
quite
comprehensively.  In the meantime I'd be most grateful if somebody could
point me in the right direction towards a solution.






Steve Huckvale
Cosworth Racing Ltd.


= = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
Subscribe: "subscribe powerh-l" in message body to majordomo@lists.swau.edu
Unsubscribe: "unsubscribe powerh-l" in message to majordomo@lists.swau.edu
powerh-l@lists.swau.edu is gatewayed one-way to bit.listserv.powerh-l
This list is closed, thus to post to the list, you must be a subscriber.