If ANS is a very long string of numeric characters it will cause this.
It will pass the pattern match (and the other tests) because each digit is
numeric, but as a whole it is not considered a number
Iirc, you had another of these around a year ago but I don't recall the
resolution
-Original
Jeff,
TEST.TRIG isn't doing very much, it's writing the same few bytes into the
same bit of memory every time.
What happens if you (re)size JUNK.TRIG to have a modulus of 5000 and change
ID on line 14 of TEST.TRIG to X.
It won't make the trigger run any faster, but it will make TEST.TRIG do a
If I've understood you correctly then you've effectively performed three
separate steps :
1. You've deleted your account(s) and restored from (straight copy) backups
2. You've deleted and reinstalled the Universe home/program account
3. You've upgraded Universe
So far as potential problems with
OP doesn't say which database, except that UniBasic suggests UniData and on
my machine it works fine in Universe but as described in UniData (5.2.3 on
RHL)!
Wide-zero is a Universe-specific feature so wouldn't affect UniData
I'm baffled - adding additional prints to the code suggests that both
Colin,
Thank you, I'd completely forgotten about Float.Precision.
Changing from (the default) setting of zero to any of the other settings
(1-4) sorts it out - although it took three passes through the HELP to
understand some of the other implications.
I'm just slightly amazed that something
Seconded.
Apologies for posting to the list instead of the moderators but let's all
consider the amount of work a small number of people put in to make this
happen, and whether we personally would be willing to step up and do it
instead of telling them how we'd like them to do it.
-Original
Dave,
I hit this with a Linux server last year and I'm hoping it's the same
problem. Pete @ Epicor UK did some research and advised that Error 39207
refers to an incorrect LANG setting on a Unix server. LANG should be set to
'C' , if it isn't you'll hit this.
My problem was that the base system
A wide0 setting of 0x3eb0 means that any difference of less than
0.01 between two values is considered insignificant and the two values
are considered equal. In your case the difference after OCONV seems to be
0.02 so the values should be considered different
You also imply that the test
Serguei,
I really don't understand your position. The illustration below is what I
would consider to be a classic example of having a single program perform
multiple functions according to data passed to it (whether the branch is
performed by an IF...ELSE, an ON ... GOSUB or a CASE statement is
The good news is that you're on this list because there are IBM'ers here who
really can (and will) resolve this, and I guess you know that.
What is really depressing is that this sort of attitude is endemic
throughout IBM (U2, DB2, AIX) in the same way that it is endemic throughout
any number of
Barry,
Search the archives for anything on wide-zero for detailed explanation, but
I'll have a go anyway.
It is all to do with binary representation of floating point numbers. If you
convert an integer from decimal to binary the resultant value is always
precise and unambiguous, the binary can
Jeff,
I think Openseq looks either for a fully specified path to the file as in :
OPENSEQ /tmp/recordname TO F.MTR.ROW ELSE
or it understands the filename, recordname construct but looks for the
filename in the VOC, as in
OPENSEQ TMPFIL, recordname to F.MTR.ROW ELSE...
where TMPFIL is a DIR
valid variable name no longer than seven characters
I think so too and I've seen systems fail for exactly that reason, but I've
just written a trivial program (see below) to test it and found that Unidata
seems to be able to differentiate between labels with up to 8 identical
characters (perhaps
Mats,
1. Sorry, my fault, hadn't properly read the previous post which stated UV
not UD
2. Looks like Unidata has gone the same way anyway, seems to cope with at
least 32 chars from v5 (or before)
Piers
---
Mats Carlid wrote:
in reply to Piers Angliss
I think he also means ensure not insure, but I daren't say I'm not
sure
Probably also about time this went [OT] if it hasn't already
P.
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] Behalf Of Bill_H
Sent: 12 October 2005 17:08
To: u2-users@listserver.u2ug.org
U2 is not a database. UniData or UniVerse ?
On UniData try HELP WHEN and look at the explanation of the ASSOCIATED
keyword
I think you're looking for
WHEN ASSOCIATED action = PCK AND act.date = (today)
but you need to make sure the file dictionary is correctly set up
also only works in ECLTYPE
Eric,
You're getting to the list Ok but it's often quiet at weekends.
Regret (well quite pleased really) that most of my work is done on Unix
servers and I can't answer the question.
Have you tried the UV documentation at
http://www-306.ibm.com/software/data/u2/pubs/library
Good luck,
Piers
MV world covers a multitude of sins
I've been using Named Common since the early 1980's, but I started on Prime
Information
So far as I know Named Common has been around as long as Unidata and
Universe - generally known as U2, which is what this list is all about
Piers
-Original
Not that it seems to matter these days but both this method and the ..-1
notation (which is an implicit form of the same thing) could be said to be
needlessly evaluating whether or not A is null 49,999 times.
An alternative would be :
A = 1
FOR I = 2 TO 5
A := @VM : I
NEXT I
I also
www.hp.com ?
It's not the easiest site to navigate but the information is there
I downloaded a PCL 5 Technical Reference manual few weeks ago, filename is
bpl13205.pdf
Piers
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] Behalf Of Mark Johnson
Sent: 08 August 2005
JayJay
It misses the point.
I'm not talking about hackers from outside, I'm talking about valid,
legitimate users with update access to the database because they have access
to the relevant application but within that application are constrained by
menus and application programs.
If I load
JayJay,
Reading between the lines I think you're saying a firewall could be a good
idea
I'm not sure that the other methods will work though. As I understand the
problem, it is that you can have a secure VB App using UniObjects on a
secure PC but if I have access to Excel on that PC,
I agree it's unlikely, but it's something that needs to be considered
I can't remember whether your branches are running PCs or dumb terminals but
if the former, would you be happy to roll out an Excel/ UniObjects reporting
tool/solution given your user population and distribution, some of whom
Pretty rare, in 20 years as a PI/U2 software/ services provider I've
experienced only one confirmed instance
This was of repeated inappropriate access by a legitimate user with a grudge
who found a back door into the database and by resetting control records
caused quite serious problems for a few
LeRoy,
Thanks for the info,
Thanks also for taking the time to read and contribute to the list
Piers
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] Behalf Of Leroy Dreyfuss
Sent: 22 May 2005 04:23
To: u2-users@listserver.u2ug.org
Subject: Re: [U2] [UV] UniObjects
Two nations, divided by a single language...(or maybe it's just my age)
UV doesn't have a scheduler as such, but there's any number of ways to
implement one - at the simplest level, Phantom a process to sleep till 3 am,
do something and Phantom itself again.
You could try using cron to do a ps
Dianne Don, thanks for the suggestions
I've got it working fine on 9.6.2 on Win 2000 now so I can play around to my
heart's content.
I take the point about upgrading, but I've got clients on 9.6 who could be
interested. Unfortunately, without maintenance an upgrade is probably too
much of an
Hi all,
Having a couple of hours to spare I decided it was time I got my head around
UniObjects but I'm having problems connecting to the database
I'm trying to connect to a UV 9.6.2 database on a Linux server and
UniObjects is returning error 39121 (Database license has expired)
I've been
!EXISTS (or similar) was definitely there in PI, I also thought it was
around in UV but I seem to remember looking for it a while back and also
drawing a blank
Piers
---
u2-users mailing list
u2-users@listserver.u2ug.org
To unsubscribe please visit http://listserver.u2ug.org/
snip
but I doubt the manager would have the appetite to spend good money just to
be able to read the very old data. /snip
snip from your website
CIBC History
The largest merger of two chartered banks in Canadian history happened on
June 1, 1961 - The Canadian Bank of Commerce (established 1867)
A very long numeric string will do this, see sample code below:
0001: ARRAY = '12345' ;* Just so array has a value
0002: VAR = STR('1',512)
0003: IF VAR MATCHES 0N THEN PRINT 'Numeric'
0004: DUM = ARRAYVAR
0005: END
RUN BP T2
Numeric
Program T2: Line 4, Nonnumeric data when numeric required.
Nice one Mike, I'll try to remember to drop you a line on 24680 - the way UK
pensions are going I'll still be working !
Piers
Tomorrow Saturday 5th March 2005 is, in the U2 / Pick calendar Day Number
13579
---
u2-users mailing list
u2-users@listserver.u2ug.org
To unsubscribe please visit
If on *nix, SAVE.LIST then quit to *nix and use the OS sort on the SAVEDLIST
record ?
Could be difficult if it's a very big list which spans multiple records (or
is that Unidata ?) but you can get round it
hth
Piers
---
u2-users mailing list
u2-users@listserver.u2ug.org
To unsubscribe please
Happy New Year all
I guess most people are already running 10 so won't be affected but for any
who are still to move from UV 9.n to 10.n, there is a change within the
compiler which made my New Year's Day very interesting.
I maintain and support a 4GL which uses the following construct
David,
Many thanks, you are absolutely right.
Epicor (UK Support) had a look and confirmed that the problem doesn't occur
at 10.1.3 (although there's no obvious reference in the release notes).
I obviously need to upgrade a little more frequently, or search the archive
more diligently ;-)
Piers
You're correct that there's an option,
put PTERM CASE NOINVERT in the LOGIN para in the VOC
The feature was implemented with the best intentions, for developers moving
between the determinedly lower-case Unix Operating System and the equally
determinedly upper-case DBMS. Some people love it,
That is the crux of the problem, there is an inherent mismatch between
numeric data stored in external format without a dictionary conversion and
numeric values entered at the command line. I think because there is no
Dictionary conversion uv is doing a string comparison not a numeric one. The
ABS
Allen,
rant
talking of achieve s (Ms$ spellchecker ?) getting full of fluff
1. What has your post got to do with PICK Assembler Language ?
2. As I understand it, the Digest is a consolidated posting so that
subscribers receive a number of postings grouped by (sometimes totally
irrelevant)
Yeah, but sometimes the response blows away the stress, and I'm told that's
good for the blood pressure
All the best,
Piers
*--
Subject: RE: [U2] What the h*** happened to subject lines ?
Sorry, I misunderstood the
Absolutely, longnames on should be automatic (and check that on in the
install automatically updates NEWACC) but imho longnames is far less of a
problem than Type1 files which are an absolute pain.
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] Behalf Of Glenn
...and how much will IBM Global Services be charging the client for that
Project Manager ?
...and I've just realised that you probably meant 35 $US not 35 ?UK
Now I see why the move into Professional Services was crucial for IBM's
future, margins like that could make even Micro$oft jealous
quote Simon Lewington
What would be great is if Unidata provided a facility to update an index in
a file programatically. /quote
If I recall correctly, Universe implements indices as Type 25 data files, so
UV could support this
Piers
---
u2-users mailing list
[EMAIL PROTECTED]
To
quoteThis is something I've managed to do on many different systems and
currently
have an open ticket with IBM (There was an old Unidata problem with
mergelists with similar message) /quote
Can you provide any details on the old problem - there may be a clue in
there
---
u2-users mailing list
At this time on a Friday evening, you can't possibly mean this weekend.
So, identify all your uv accounts over the next week and, for each one, scan
the VOC looking for any references in F2 to the TRANS file. Make a note of
them and, when you make the move, go back and change each one to point to
Brad,
forgive me if I've missed the point completely, but DUP and LOAD aren't vi.
If DUPenter worked you're using AE, Unidata's Alternate Editor, type
HELP to get documentation.
If DUP pasted another copy of the current line at the end of it then you are
using vi, but I don't think LOAD would
IMHO.
They are very similar, with hundreds of differences - whether any are
significant to you depends on where you're coming from and what you want to
do.
They are both M/value (Pick-like, post-relational, NFNN ...) multi-user
databases running on *nix or Windows servers.
They have an
To: [EMAIL PROTECTED]
Subject: Re: [U2] Comparison Unidata and Universe
Piers Angliss wrote:
IMHO.
They are very similar, with hundreds of differences - whether any are
significant to you depends on where you're coming from and what you want to
do.
Where are you coming from and where do you
Adrian,
I'm surprised you haven't seen it (not that I'd wish the subsequent
reconciliation and explanation on anyone)
If you're interested here's an example :
ED BP TWZ
6 lines long.
: P
0001: NET = 294987.31
0002: VAT = 51622.78
0003: TOT = NET + VAT
0004: IF TOT EQ 346610.09 THEN PRINT
Claus,
Fair comment and you're absolutely right that this will avoid the problem.
It occurs (I think) because although you can express every integer within
the allowable range precisely in binary, you cannot do the same with every
fraction of an integer (what's the word I'm looking for ?).
So, if
Ray,
A detailed and complete explanation as always, thanks
I think I've finally understood it !
IEEE standards say zero is 2.91E-11 , which as a fraction is roughly 1 /
pwr(2,35)
So, the binary representation of IEEE zero has an exponent of -35
If you bias this by 1023 you get 988 or
...and that's the problem.
The answer is that every so often if you write code like mine which adds /
subtracts numbers in external format (ie with decimal points) universe will
discover an infinitesimal difference between two numbers that to our eyes
(and even in truth) are absolutely identical.
Universe 9.n on Unix :-
Does anybody have any advice or horror stories related to configuring WIDE0
?
I have experienced problems in the past on a couple of client sites running
the default 0x3dc0 value and have established, by experiment, that a
value of 0x3dd0 would have solved the
Personally I use it all the time (except I'm even sloppier and I drop the
DO )
It's been around for as long as I can remember so I think it just isn't
documented
I find it pretty readable but I take your point on using undocumented
features - boy, will I be in trouble if IBM ever change it.
Piers
snip
The CONVERT statement does not change the way you think.
/snip
Shame - that really would be a unique selling point for U2 !
---
u2-users mailing list
[EMAIL PROTECTED]
To unsubscribe please visit http://listserver.u2ug.org/
snip
The Alert Centre, Englewood CO USA, 1993 - 1996
The Alert Centre (yes they really spelled it that way) was an security
\snip
Sorry - how else could they have spelt Alert ?
---
u2-users mailing list
[EMAIL PROTECTED]
http://www.u2ug.org/listinfo/u2-users
55 matches
Mail list logo
