That's it, yes. I've used this to send messages between processes. But
using them can be less than straightforward. The GCI manual refers to
UVCreatePipe and related functions, that look more likely to be more
robust.
/Andre
--
Date: Tue, 3 Jan 2006 11:09:48
Happy New Year.
I am having trouble with the openseq function in UniData. We are
operation in Pick mode.
Openseq can't find the specified file at path,filename. Is there an
assumed prefix?
Here is the code frag that is failing.
MTR.FILE.PATH='/tmp'
OPENSEQ MTR.FILE.PATH, MTR.REC TO
Jeff,
AFAIR OpenSeq on UniData looks for either a path name, or if you specify an
item id it looks for an item in a directory file (not a path).
So you would need to either
A) OPENSEQ /tmp/ : MTR.REC TO F.MTR.ROW.FILE ...
Or
B) create a file pointer to /tmp (e.g. named TMP) and then
Is it a permissions problem?
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] Behalf Of Jeff Powell
Sent: Wednesday, January 04, 2006 8:31 AM
To: u2 users group
Subject: [U2] openseq question
Happy New Year.
I am having trouble with the openseq function in UniData.
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
No. /tmp is an absolute path.
On Wed, 2006-01-04 at 09:47 -0500, Koser, Mike wrote:
Jeff
Is '/tmp' in the VOC?
Mike
-Original Message-
From: Jeff Powell [mailto:[EMAIL PROTECTED]
Sent: Wednesday, January 04, 2006 9:31 AM
To: u2 users group
Subject: [U2] openseq question
Happy new projects in the New Year (may we each have all we can handle,
and may they all pay).
I'm still trying to wrap my mind around the UniVerse implementation of
Java.
I need to trigger a call, from UV, to a Java API, sending Credit Card
data to our processor.
From what I can read, the
Your syntax is a little off. When you use the seq.file.name, record.id
syntax then seq.file.name has to be a dir type file. You will have to
either create a voc entry for /tmp, eg.:
TMP.DIR
001 DIR
002 /tmp
003 D_VOC
And then you should be able to: MTR.FILE.PATH = TMP.DIR
Or you can OPENSEQ
You need to have MTR.FILE.PATH open as a file handle, first. You can do
one of the following:
MTR.FILE.PATH='/tmp'
OPEN MTR.FILE.PATH TO MTR.FILE ELSE STOP
MTR.REC = 'RECID'
OPENSEQ MTR.FILE,MTR.REC TO F.MTR.ROW.FILE ELSE
[what you already have]
END
Or open the full path directly:
I've actually had problems with various versions of Unidata and
CLOSESEQ releasing the file when opening a full path. Opening under a
VOC DIR-pointer has never exhibited that problem, so I'd recommend the
VOC pointer approach over using a full path.
-Kevin
[EMAIL PROTECTED]
Just to clarify... You want to move information FROM Universe TO a
Java app? What interfaces (communication interfaces, not Java
interfaces) are provided in the Java app?
-Kevin
[EMAIL PROTECTED]
http://www.PrecisOnline.com
---
u2-users mailing list
u2-users@listserver.u2ug.org
To
Several years ago I created a little utility to open sequential files as
needed. It's pretty robust; hasn't given me much trouble over the years.
Like Kevin, I've had problems with the full path version of openseq, so my
utility converts full path mode to VOC pointer mode and creates the VOC
Wendy wrote:
IBM provides an implementation of the JDBC specification, (a JDBC
driver,) as well as a proprietary Java API known as UniObjects for
Java.
Yes, I was reading the docs on it, but no code samples there, so things
still a bit fuzzy.
If you mean the the JDBC CallableStatement, you would
I've always found it helpful to open the file/record in 'regular' mode first
to find out if the directory entry worked or possibly to write a null record
if it didn't.
I like to have a graph display, and this way I can tell how many 'records'
(or AMCs in this case) I'm going to be processing so I
Hi Kevin
Do you recall which OS this was on? I don't believe I've run into this
problem but I'm currently using Windows (2000 and 2003) but I've also
been on RS6000 and DG/UX. All of this is with UV.
Thanks.
BobW
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED]
Thank you for all your helpful suggestions.
I am now looking at creating a VOC pointer as Kevin suggests below.
Actually I'll need to create two pointers one for the TEST account and
another for the LIVE account.
I had originally tried the PATH:FILE syntax before I found out that the
file needed
Do you recall which OS this was on? I don't believe I've run into
this problem but I'm currently using Windows (2000 and 2003) but I've
also been on RS6000 and DG/UX. All of this is with UV.
Unidata on Windows, as I recall.
-Kevin
[EMAIL PROTECTED]
http://www.PrecisOnline.com
---
u2-users
Jeff;
I use a VOC pointer for exactly this reason. It doesn't matter which
account/where the directory is - I just need the VOC pointer that points
to it. It makes it really easy to transfer from TEST to LIVE.
Also, as Tim pointed out, you don't need the file to exist to open it.
In fact, I have
Unidata on Unix has this issue, too (at least the DEC Tru64 flavor) -- it holds
onto the file lock after CLOSESEQ of a file OPENSEQed with the Unix path.
Don't know if it's been fixed in newer UD releases, but not likely a priority
since OPENSEQ of a VOC pointer doesn't have the problem.
Actually I'll need to create two pointers one for the TEST account
and another for the LIVE account.
This is one of the great features of the VOC pointer; the LIVE account
can point to the directory it needs and the TEST account can point to
the directory it needs and the program doesn't have to
You'd be better off using the socket interface for most communication - more
robust, and not tied to the single platform (pipes are internal to a system
and don't go outside).
See the UV 10 doco - sample programs available from your friendly, local U2
support desk.
JayJay
---
u2-users
One option is to put a C wrapper around the Java API and link it in via the
GCI interface. However lots of companies provide a C API version as well so
maybe you can link directly?
Otherwise:
1. MQ series (MQ interface via AMI) - not available on ALL platforms, though
a newer version will be
On 1/4/06, Baker Hughes [EMAIL PROTECTED] wrote:
Yes, I was reading the docs on it, but no code samples there, so things
still a bit fuzzy.
Look in (the equivalent of) c:\ibm\unidk\uojsdk and c:\ibm\unidk\jdbc
for samples.
I'm trying to call the Live Processor (from Retail Decisions) Java
Has anyone tried to running Universe on a Windows 2003 Server (32 bit) not
running on an Intel chip but a AMD Operton Chip, I don't see it as a option
on their platform matrix, hoping someone has tried this.
Alternatively has anyone tried running on Intel/Xeon chip.
Thanks
Peter Wezenaar
I've run Unidata on an NT 4.0 and an AMD 450 about 6 years ago...worked
fine. I still have it setting on a shelf in case I ever need to resurrect
it... . .
Allen E. Elwood www.tortillafc.com
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] Behalf Of Peter Wezenaar
We're currently running Unidata 6.0 and I have been trying unsuccessfully to
call a Unidata subroutine (Stored procedure) via SQL. There isn't a whole
lot of documentation on how to call a stored procedure. I'm using the
following syntax: {CALL ROUTINE_NAME('','1234567')}
The name lookup
Peter,
-Original Message-
From: [EMAIL PROTECTED] On Behalf Of Peter Wezenaar
Sent: Thursday, 5 January 2006 12:11
To: u2-users@listserver.u2ug.org
Subject: [U2] Running Universe on AMD processor
Has anyone tried to running Universe on a Windows 2003 Server
(32 bit) not running on
The Opteron can run 32 and 64 bit applications. I have Win2K running on a
155 chip and the performance/price blows away Xeons.
Glen
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] Behalf Of Peter Wezenaar
Sent: Wednesday, January 04, 2006 6:11 PM
To:
Thanks again to everyone for all their help. I'm almost there.
I want to use the VOC pointer so I created the pointers in my Live and
Test VOC files, logged out and back in to ensure I was reading VOC but
I'm still not meeting with any success. When I attempt to reference the
pointer I get a
Jeff, if the VOC entry is MFG.DATA.IPS.TMP then the OPEN is:
OPEN MFG.DATA.IPS.TMP, MTR.REC TO ...
Not the concatenated thing you had in your post.
-K
---
u2-users mailing list
u2-users@listserver.u2ug.org
To unsubscribe please visit http://listserver.u2ug.org/
Jeff Powell wrote:
Thanks again to everyone for all their help. I'm almost
there.
I want to use the VOC pointer so I created the pointers in
my Live and
Test VOC files, logged out and back in to ensure I was
reading VOC but
I'm still not meeting with any success. When I attempt to
reference
Allen E. Elwood wrote on 01/04/2006 01:29:47 PM:
I've always found it helpful to open the file/record in 'regular' mode
first
to find out if the directory entry worked or possibly to write a null
record
if it didn't.
So, if the O/S-level file is multi-megabytes, you're reading the whole
Further to this, you could also use the !GET.PATHNAME subroutine to get
the path and filename. A quick way to check rather than the expense of a
DB level OPEN.
Regards
---
** _/ ** David Logan
*** _/
Happy New Year,
A current project involves printing a number string in its alphabetic
description e.g. 1 = one, 3 = three, 11 = eleven etc etc. The precise task
is printing a label of book information and in this case the label is
edition e.g. first/second.
The universe basic manual does not
Goo'day, David,
Dick Pick's gift to the world:
At 15:24 05/01/06 +1100, you wrote:
Happy New Year,
A current project involves printing a number string in its alphabetic
description e.g. 1 = one, 3 = three, 11 = eleven etc etc. The precise task
is printing a label of book information and in
How high are you looking to go? Seems a simple CASE might do the
trick if the numbers are reasonably small. If you have to get into
some big numbers, split the number into triads and parse the triads
independently. That is, xxx is the hundreds triad, xxx,xxx is the
thousands triad, xxx,xxx,xxx
We're running UV 10 on a quad Xeon box with Windows 2000 server - without
any problems at all. Windows 2003 shouldn't be a problem either - I
originally tested our server with this, but fell back to W2k because we
implemented _just_ after W2k3 became available, and I didn't want our
enterprise
37 matches
Mail list logo
