Hi,
This is with relating to 2.0.9.15 IF2
The good news is Cube Dependancy has been fixed but there seems to a be another problem which is with support at the moment.
I have found the new issue, after some more investigation, around using DIMIX in a rule formula and a process picking up the rule calculation value
Problem Summary
a) Have a alias rule on the }ElementAttribute Cube
b) Have a DIMIX rule on in a calculation
c) Retrieve that calculation in a process with a CellgetN
If those three things are in play then the process runs really slowly.
If you dont use DIMIX in your rule calculation or dont have a alias rule on your }ElementAttribute Cube then the process works fine as normal in 2.0.9.11
------------------------------
Herman Moller
------------------------------
Original Message:
Sent: Fri January 27, 2023 06:10 AM
From: Steven Rowe
Subject: 2.0.9.15 issue : ATTRS rules in }ElementAttribute cube causes a TI to load data slowly
Hi Herman,
Happy New Year! Hope all is well with you? (baring the obvious issues highlighted here!)
As a cube dependacy is being built AddCubeDependency('S_Consolidated','}ElementAttributes_Account');
I'm assuming there is a rule in S_Consolidated that makes reference to the Account Attributes?
Is it the exact ruled attribute that is referenced by the the rules in S_Consolidated?
Or is it just the fact that are any ruled attributes in the cube that cause the issue, i.e. the rule sheet only references value based attributes?
Just trying to understand the detail so we can avoid the issue with any of customers.
Cheers,
------------------------------
Steven Rowe
Original Message:
Sent: Thu January 26, 2023 05:11 AM
From: Herman Moller
Subject: 2.0.9.15 issue : ATTRS rules in }ElementAttribute cube causes a TI to load data slowly
Morning, Unfortunately we have found even through performance has improved in some areas this is still performances issue in comparison to previous versions of Planning Analytics.
------------------------------
Herman Moller
Original Message:
Sent: Tue January 17, 2023 06:07 AM
From: Herman Moller
Subject: 2.0.9.15 issue : ATTRS rules in }ElementAttribute cube causes a TI to load data slowly
IBM Support have come back to us. They have a fix - 2.0.9.15 IF2 which resolves the issue. We are currently testing the fix pack now.
------------------------------
Herman Moller
Original Message:
Sent: Tue November 15, 2022 05:18 AM
From: Herman Moller
Subject: 2.0.9.15 issue : ATTRS rules in }ElementAttribute cube causes a TI to load data slowly
Hi Tyler,
Thanks for the suggest it was actually one of the first things I tried but it still to have the same issue. ATTRS and DB referencing the }ElementAttribute cube made the TI run much slower.
------------------------------
Herman Moller
Original Message:
Sent: Mon November 14, 2022 11:06 AM
From: Tyler King
Subject: 2.0.9.15 issue : ATTRS rules in }ElementAttribute cube causes a TI to load data slowly
I don't know why the Cube Dependency is affecting the TI run times, but try updating the ATTRS function to A DB function in the Rule if possible. The DB function generally performs better than ATTR functions in rules. The only caveat is if you are referencing an alias with ATTRS. ATTRS will return the element principle name if the alias is blank in the }ElementAttributes_ cube, but DB will return a blank value if it's not populated. Just something to keep in mind.
------------------------------
Tyler King
Original Message:
Sent: Fri November 11, 2022 04:24 AM
From: Herman Moller
Subject: 2.0.9.15 issue : ATTRS rules in }ElementAttribute cube causes a TI to load data slowly
HI,
We are looking to upgrade our Planning Analytics server to the latest version 2.0.9.15 from 2.0.9.11 and have notice some significant issues in transfer data from a cube to cube transfer in both versions 2.0.9.13 and 2.0.9.15.
We have 70 lines of data from One cube which normally takes seconds but when we upgrade to the new version it is significantly slower
I have isolated the issue to being :
1) Have a rule in a element attribute cube that uses the ATTRS function
2) Have cube dependencies run on start up.
I have found if I don't run cube dependencies on start-up and run the process it runs fine.
When I activate cube dependencies it slows the TI down significantly.
I have opened a case with IBM but has anyone else experienced this problem?
Regards
Herman
------------------------------
Herman Moller
------------------------------
#PlanningAnalyticswithWatson