Linking to a subfile (was Subphyla question)

[mao@krifapost.dk]

Amen to that!

>On every other list I've ever subscribed to (and that's quite a few - maybe
>I should get a life), the Reply-To address in the mail header has been set
>to the *list address*.  On this list it's the *originator address*, which
>makes it easy to turn a discussion off-list accidentally.
>
>Is there any chance of making this list like all the others?
>
>Chris Walker
>Tate & Lyle
>London


> -----Original Message-----
> From: Robert J.M. Edis [mailto:robert.edis@creatcomp.com]
> Sent: Wednesday, March 01, 2000 7:34 PM
> To: 'powerh-l@list.swau.edu'
> Subject: RE: Linking to a subfile (was Subphyla question)
>
>
> Sorry Richard
>
> I originally wrote to Matt privately as I didn't think it was
> of general
> interest, but it got interesting when Bob waded in.
>
> Blue
>
> -----Original Message-----
> From: Richard Witkopp
> To: 'powerh-l@list.swau.edu'
> Sent: 3/1/00 1:19 PM
> Subject: RE: Linking to a subfile (was Subphyla question)
>
> Wow, I never received any of this discourse except maybe the
> first message. I wonder if the problem is with the listserver
> or our mail system here.
>
> > G'day Matt, Bob
> >
> > Actually you may have hit the nail on the head Bob.  The
> > report I mentioned
> > to Matt did have a more complex ACCESS than I intimated.  It
> > was more along
> > the lines of
> >
> > ACCESS table1 &
> >   LINK TO table2 OPTIONAL &
> >   LINK TO RECORD 0 OF *subfile
> >
> > SELECT IF ....
> >
> > This took about 45 minutes to run even though only a few
> > thousand records
> > were involved.
> >
> > By changing it to
> >
> > ACCESS table1 &
> >   LINK TO table2 OPTIONAL
> > CHOOSE SYSTEMVALUE ...
> > SELECT IF ...
> >
> > reduced the run time down to about 5 minutes even though the
> > same number of
> > records were read from the primary file.
> >
> > Blue
> >
> > -----Original Message-----
> > From: Deskin, Bob
> > To: Ohmes, Matt; 'Robert J.M. Edis'
> > Sent: 3/1/00 6:40 AM
> > Subject: RE: Subphyla question
> >
> > Theoretically (love that) a link to a specific record of a
> > subfile (via
> > LINK
> > TO RECORD) should be treated like a link to a unique index.
> As long as
> > the
> > index value doesn't change, QUIZ doesn't read the record over
> > because it
> > "knows" that the record already in the buffer is the correct
> > one. That's
> > the
> > simple case. However, if you throw in SELECTs and SELECT
> file IFs and
> > OPTIONALs, the situation gets more complicated. There are
> cases where
> > the
> > input buffers are cleared because otherwise a DEFINE could be
> > evaluated
> > using data belonging to another complex.
> >
> > The way to test this is something like
> >
> > ACCESS A LINK TO B LINK TO RECORD 0/1 OF C
> > GO
> >
> > The 0 or 1 to get the first record depends on the platform.
> C can be a
> > direct file or a subfile, in which case put an asterisk on
> the front.
> >
> > There should be very little difference between using C or
> not using C
> > (ACCESS A LINK TO B). But of course you have to run something
> > that takes
> > a
> > reasonable amount of time because there will be file open overhead.
> >
> > As Matt suggested, this used to work but probably hasn't been
> > checked in
> > a
> > long time. In fact I think it's documented that way, at least
> > somewhere
> > in
> > all the PH manuals on my shelves :-)
> >
> > Bob
> >
> > -----Original Message-----
> > From: Ohmes, Matt
> > Sent: February 29, 2000 10:12 PM
> > To: 'Robert J.M. Edis'
> > Cc: Deskin, Bob
> > Subject: RE: Subphyla question
> >
> >
> > Hmmm... Well, to tell you the truth, I've never specifically
> > tested it,
> > but
> > I'd heard this is the way it worked.  And I've never had the type of
> > performance problem you describe. I've used this technique
> a lot over
> > the
> > years and never saw the type of performance you described, but I've
> > never
> > specifically tried to test it.
> >
> > Now I must admit, I rarely worked on a UNIX box. My HP
> > experience almost
> > exclusively on HP3000, but I've also used the technique
> extensively on
> > VAX.
> >
> > Also, were you accessing a relational table then linking to
> > the subfile?
> > If
> > so, the performance enhancement is almost certainly
> > attributable to the
> > CHOOSE vs. a SELECT and not the Variable per se.
> >
> > Do I understand correctly the way we read a single record
> subfile Bob?
> > Matt
> >
> > -----Original Message-----
> > From: Robert J.M. Edis [mailto:robert.edis@creatcomp.com]
> > Sent: Tuesday, February 29, 2000 5:29 PM
> > To: 'Ohmes, Matt '
> > Subject: RE: Subphyla question
> >
> >
> > G'day Matt
> >
> > I have to challenge you on what you said about "Quiz will read the
> > subfile
> > record once".  I 'fixed' a report for a client last year on a
> > HP/UX box
> > where a report was taking 45 minutes to run.  This report was
> > reading a
> > table and linking to record 1 of a subfile.  It read the subfile for
> > each
> > occrrence on the primary file.
> >
> > I replaced the logic using an environment variable and a
> CHOOSE on the
> > primary file instead.  This reduced the run time to only 5 minutes.
> >
> > Blue
> >
> > -----Original Message-----
> > From: Ohmes, Matt
> > To: 'Scott, Valerie'; POWERH-L@lists.swau.edu
> > Sent: 2/29/00 5:48 PM
> > Subject: RE: Subphyla question
> >
> > Actually you can link to the subfile from your data file. Try this:
> >
> > ACCESS datafile LINK TO RECORD 1 OF *subfile
> >
> > SELECT IF dateitem of subfile = dateitem of datafile
> >
> > Quiz will read the subfile record once and keep it in memory
> > through the
> > entire run, so it should be quite effecient.
> >
> > You didn't say what machine you are on, but the above works
> > for VAX and
> > UNIX
> > (I believe). If you are on an HP3000, substitute a zero for
> the "1" in
> > the
> > ACCESS statement.
> >
> > Matt
> >
> > Matt.Ohmes@Cognos.Com
> > Cognos Corporation
> > 909 E. Las Colinas Blvd.
> > Suite 1900
> > Irving, TX  75039
> > 214-259-6200
> > "Matt doesn't really know anything.  He just likes to
> > pontificate a lot.
> > We refuse to acknowledge that he works for Cognos or that
> we have ever
> > met him or anyone with whom he's ever been associated.
> Don't lend him
> > money and don't let him talk to your sister!" ;-)
> >
> > -----Original Message-----
> > From: Scott, Valerie [mailto:vscott@solanocounty.com]
> > Sent: Tuesday, February 29, 2000 4:30 PM
> > To: POWERH-L@lists.swau.edu
> > Subject: Subphyla question
> >
> >
> > How can I avoid needing a parm prompt entry in all 4 passes
> > of a report?
> > I
> > am unable to pass the user date range beyond the first report.
> >
> > I am writing a Quiz program that runs 4 passes of data.
> > Each pass for data require the same user entered Date range.
> > I created a one record subfile(a) to accept the PARM PROMPTED date
> > value.
> > On the first report, I use a USE to run the subfile(a),
> prompting for
> > the
> > date.
> > The remaining reports access other files and append to a separate
> > subfile,
> > which at the end is sorted and reported by type of record.
> > I don't believe I can link to my 'date' subfile, because
> the data file
> > is
> > driving the reporting.
> > Once the date range is captured, is there a way to hold the dates,
> > throughout all four passes?
> >
> > Thank You,
> > Valerie
> > Solano County
> >
> >
> >
> >
> > Thank you,
> >
> > Valerie Scott
> >
> >
> > = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
> > = = = = =
> > = =
> > Subscribe: "subscribe powerh-l" in message body to
> > majordomo@lists.swau.edu
> > Unsubscribe: "unsubscribe powerh-l" in message to
> > majordomo@lists.swau.edu
> > This list is closed, thus to post to the list, you must be a
> > subscriber.
> > = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
> > = = = = =
> > = =
> > Subscribe: "subscribe powerh-l" in message body to
> > majordomo@lists.swau.edu
> > Unsubscribe: "unsubscribe powerh-l" in message to
> > majordomo@lists.swau.edu
> > This list is closed, thus to post to the list, you must be a
> > subscriber.
> > = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
> > = = = = = = =
> > Subscribe: "subscribe powerh-l" in message body to
> > majordomo@lists.swau.edu
> > Unsubscribe: "unsubscribe powerh-l" in message to
> > majordomo@lists.swau.edu
> > This list is closed, thus to post to the list, you must be a
> > subscriber.
> >
> = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
> = = = = =
> = =
> Subscribe: "subscribe powerh-l" in message body to
> majordomo@lists.swau.edu
> Unsubscribe: "unsubscribe powerh-l" in message to
> majordomo@lists.swau.edu
> This list is closed, thus to post to the list, you must be a
> subscriber.
> = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
> = = = = = = =
> Subscribe: "subscribe powerh-l" in message body to
> majordomo@lists.swau.edu
> Unsubscribe: "unsubscribe powerh-l" in message to
> majordomo@lists.swau.edu
> This list is closed, thus to post to the list, you must be a
> subscriber.
>
= = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
Subscribe: "subscribe powerh-l" in message body to majordomo@lists.swau.edu
Unsubscribe: "unsubscribe powerh-l" in message to majordomo@lists.swau.edu
This list is closed, thus to post to the list, you must be a subscriber.


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