We use PRO-IV 4.0 here with Oracle 7.3.4 a lot of years..
I have a test with Oracle 8.1.6 and 8.1.7, using 4.0
link with older libraries (of 7.3.4) and all was OK..
But when I use Oracle 9.2, all are OK, minus decimal
part of numbers..
In SQL*PLUS come with 7.3.4 (older. older. older), All be OK..
But in PRO-IV, we dont see decimal part of number...
Exemple-> 139.50 in Oracle, 139.00 in Data Browser in PRO-IV
but if number or digits is lower, all be OK..
Exemple-> 13.55 in Oracle, 13.5 in Data Browser in PRO-IV
I Know that PRO-IV 4.0 don't have suport of Oracle 9, but
don't have suport of Oracle 8i too, and works!!
(lucky ??)
Any sugestion ??
Thank You very much
Click the link below to see the new game I'm developing!
Pro-IV 4.0 Truncate decimal numbers when Oracle
Started by
Guest_marcus minervino_*
, Aug 01 2002 05:49 PM
19 replies to this topic
#2
Guest_marcus minervino_*
Guest_marcus minervino_*
-
- Guests
Posted 01 August 2002 - 05:53 PM
One comment.. Oracle Client have been stay stable for
a lot of versions...
I dont understaind why dont work well in Oracle 9.2, if
SQL*PLUS Release 3.3.2.0.0, distributed with Oracle7 Server Release 7.3.4.3.0, works fine, and PRO-IV link with
libraries of this version have this problem..
Maybe some parameter ??? (NLS_NUMERIC_CHARACTERS??)
Thank You
a lot of versions...
I dont understaind why dont work well in Oracle 9.2, if
SQL*PLUS Release 3.3.2.0.0, distributed with Oracle7 Server Release 7.3.4.3.0, works fine, and PRO-IV link with
libraries of this version have this problem..
Maybe some parameter ??? (NLS_NUMERIC_CHARACTERS??)
Thank You
#3
Richard Bassett
-
- Members
-
- 707 posts
ProIV Guru
- Gender:Not Telling
- Location:Rural France
Posted 01 August 2002 - 06:07 PM
Assuming you're still using the 7.3.4 client-side libraries linked with ProIV it's quite possible they don't work 100%any more with a 9.2 server!
It is possible that the reason SQL*plus seems to work is because it uses a different subset of the OCI than ProIV.
For example maybe SQL*Plus binds the number columns in a different way.
Do you know if Oracle officially support access to a V9.2 server from a V7.3 OCI client? If not then it's probably not a good idea to use it anyway
It might be possible to change the ProIV link script yourself to link your old ProIV with Oracle V9 OCI libraries - or it might not if the OCI has changed too much.
It is possible that the reason SQL*plus seems to work is because it uses a different subset of the OCI than ProIV.
For example maybe SQL*Plus binds the number columns in a different way.
Do you know if Oracle officially support access to a V9.2 server from a V7.3 OCI client? If not then it's probably not a good idea to use it anyway
It might be possible to change the ProIV link script yourself to link your old ProIV with Oracle V9 OCI libraries - or it might not if the OCI has changed too much.
Nothing's as simple as you think
#7
Dan Shannon
-
- Members
-
- 374 posts
ProIV Guru
- Gender:Male
- Location:Australia
Posted 01 August 2002 - 09:14 PM
We are experiencing precisely the problem described with PRO-IV v5.0r109 on Solaris, with the 8i and 9i 32-bit and 64-bit clients talking to a 9i (release2) Oracle database. We do not encounter the problem with the 8i database.
PRO-IV are apparently working on a fix...
Dan Shannon
PRO-IV are apparently working on a fix...
Dan Shannon
#8
Guest_marcus minervino_*
Guest_marcus minervino_*
-
- Guests
Posted 01 August 2002 - 09:36 PM
We dont sold PRO-IV 5.. but we make a test with it on
HP-UX 11.11, link 5.0 on libraries of 8.1.7.4 (32 bits), and
talk with Oracle 9.2.0.1 (64bits) on same machine and
don't found this problem...
But I will repeat my tests...
Thank You
HP-UX 11.11, link 5.0 on libraries of 8.1.7.4 (32 bits), and
talk with Oracle 9.2.0.1 (64bits) on same machine and
don't found this problem...
But I will repeat my tests...
Thank You
#9
Richard Bassett
-
- Members
-
- 707 posts
ProIV Guru
- Gender:Not Telling
- Location:Rural France
Posted 02 August 2002 - 08:44 AM
Ah, so maybe it is down to ProIV then...
As a matter of interest Dan, do you know if it exhibits this behaviour using both the NUMBER and ORANUM external types or only with NUMBER?
Thanks, Richard
As a matter of interest Dan, do you know if it exhibits this behaviour using both the NUMBER and ORANUM external types or only with NUMBER?
Thanks, Richard
Nothing's as simple as you think
#10
Dan Shannon
-
- Members
-
- 374 posts
ProIV Guru
- Gender:Male
- Location:Australia
Posted 02 August 2002 - 10:09 AM
Hi Richard
The ORANUM and NUMBER problems that we know about are *different* problems - one prevents reliable comparison between Oracle file variables and other (scratch) vars - this is the ORANUM issue - and the other simply doesn't return the correct values to the PRO-IV kernel.
APPARENTLY there's a fix available for the ORANUM issue under Solaris 2.8/Oracle 9i/PRO-IV v5.0 - I'm hoping to get it this weekend. Will let the forum know as to testing progress.
Cheers
Dan Shannon
The ORANUM and NUMBER problems that we know about are *different* problems - one prevents reliable comparison between Oracle file variables and other (scratch) vars - this is the ORANUM issue - and the other simply doesn't return the correct values to the PRO-IV kernel.
APPARENTLY there's a fix available for the ORANUM issue under Solaris 2.8/Oracle 9i/PRO-IV v5.0 - I'm hoping to get it this weekend. Will let the forum know as to testing progress.
Cheers
Dan Shannon
#11
Dan Shannon
-
- Members
-
- 374 posts
ProIV Guru
- Gender:Male
- Location:Australia
Posted 07 August 2002 - 01:12 PM
Just so that y'all know -
We got a patch for the ORANUM problem from PRO-IV for v5.0r109 (on Solaris) today - and it seems to work so far (in our development environments at least)
My general opinion is that ORANUM is a far preferable technical solution to handling Oracle number columns than NUMBER, so I would suggest that anyone encountering the problems discussed in this thread (and other, older ones) should badger PRO-IV support into providing the same fix to them.
Hopefully there'll be a new kernel soon which will include this patch too
Cheers
Dan Shannon
We got a patch for the ORANUM problem from PRO-IV for v5.0r109 (on Solaris) today - and it seems to work so far (in our development environments at least)
My general opinion is that ORANUM is a far preferable technical solution to handling Oracle number columns than NUMBER, so I would suggest that anyone encountering the problems discussed in this thread (and other, older ones) should badger PRO-IV support into providing the same fix to them.
Hopefully there'll be a new kernel soon which will include this patch too
Cheers
Dan Shannon
#12
Marcus Minervino
-
- Members
-
- 8 posts
Newbie
- Gender:Male
- Location:Bsb, Brazil
Posted 21 August 2002 - 10:20 AM
Dan. You comment that ORANUM is better than NUMBER.
Questions:
1) Where in www.proiv.com restricted area I can get a fix
to this problem..I lock for it and dont found this
2) If I get this fix and change all of my external definitions from NUMBER to ORANUM, All will work fine???
Any change will be need in programs ???
Thanks Dan
Questions:
1) Where in www.proiv.com restricted area I can get a fix
to this problem..I lock for it and dont found this
2) If I get this fix and change all of my external definitions from NUMBER to ORANUM, All will work fine???
Any change will be need in programs ???
Thanks Dan
Reply to this topic
0 user(s) are reading this topic
0 members, 0 guests, 0 anonymous users
Click the link below to see the new game I'm developing!
