MDX formulae reduces application performance

Hi,
I have few MDX formulae in my account dimension which is built as per our reporting requirements. But unfortunately having these formulae in greately affecting the performance. It takes 7- 8 minutes to generate a report if I have formulae.
Can anyone tell me is there anyway I can increase the performance of the application without removing MDX formulae?
Thanks
Sharath

Ok. Let me explain like this:
Net Income is the highest level made up of below children and grand children.
Memeber--Hlevel--Parent--
Description -
EAT--H5--
(Earnings After Tax)
Tax -
H4----
EAT
EBT--H4EAT--
(Earnings Before Tax)
Depreciation--H3--
EBT
Interest--H3--
EBT
EBIDT--H3EBT--
(Earnings before Interest, Depreciation and Tax)
GP--H2EBIDT--
(Gross Profit)
REVENUE--H1--
GP
COST--H1--
GP
In the above structure if EBIDT is a calculated member then I should add formula to both EBT and EAT as per MDX rule even though I can have this done through roll-ups. Other wise EBT and EAT will just return the value of EBIDT and ignores Interest, Depreciation and tax.
Without formula it takes 25-30 sec. and with formula it takes 500-600 sec. to get the numbers from the report.
Edited by: Raghavendra Sharath on Mar 20, 2009 6:03 PM

Similar Messages

Urgent - Oracle Applications Performance - Please Help me!

Hi folks!
I´m having a lot of performance troubles in Oracle applications, and the dba and the network analyst can´t help me. I´m having this trouble since 2 weeks, and nobody can explain why. The system is too slow and if i try a simple select in database it takes a lot of time, i think (i think ok?) there are som database troubles...
I saw some docs
1. System Mgmt White Paper http://www.oracle.com/appsnet/technology/managing/collateral/wp_managing11i.pdf
2. System Mgmt PPT http://www.oracle.com/pls/oow/oow_user.download?p_event_id=15&p_file=P39948.zip
3. Reducing 11i Downtime PPT http://www.oracle.com/pls/oow/oow_user.download?p_event_id=15&p_file=P39947.zip
4. Performance and Scalability site : There are a couple of excellent presentations and white papers which will give you the right way to do performance tuning. http://www.oracle.com/appsnet/technology/performance/content.html
but i need something more specifical like
simple performance tests.
Best Regards!
Filipe
[email protected]

Hi
Check this one.
http://www.appsworld2004.com/scps/controller/catalog
Search for item 1066. This is a presentation on "Performance Tuning Users Tips and Techniques" by Ahmed Alomari, Applications Performance Group, Oracle Corporation. You may need to register and then login.
If you cant access this presentation, let me know, I can mail it to you as well.
There was a similar presentation of 2002 Appsworld as well. Am not able to locate the link yet.
Best Wishes
Vinod Subramanian

Unit Of Order reducing the performance after some threshold limit

Hi,
Our application is doing netting of trades.
We are using UOO for posing messages to JMS queues appended to it.
From this message, we are extracting the trade info and persisting into DB.
When MDB is listening to that JMS queue, application is performing well when the no. of messages in the queue is <5000.
when this number increases lets say 20000, our application is performing very badly.
We are using 64 instances of MDB connecting to queue.
What are the parameters reduce the performance of the MDB consuming the messages?
Can you please help me on this?
Thanks,
Ravikumar

An increasing backlog is a sign that your MDBs are not keeping up with your senders.
See:
Tuning WebLogic JMS
http://docs.oracle.com/cd/E17904_01/web.1111/e13814/jmstuning.htm#PERFM294
and
Tuning Message-Driven Beans
http://docs.oracle.com/cd/E17904_01/web.1111/e13814/mdbtuning.htm#PERFM271

Does sub program reduces the performance

hi all,
suppose that i have two solutions to increase the salaries of the employees for a table
method 1:
create or replace function inc_sal (pempno emp.empno%type) return integer
is
vsal emp.sal%type;
begin
select sal into vsal from emp where empno=pempno;
vsal:=vsal+vsal*0.30;
return(vsal);
end;
sql> select sal, inc_sal(empno) from emp;
and
method 2:
sql> select sal, sal*0.30 as inc_sal from emp;
now in the first method, for each record of the table the function is called and collect the returned value...
but for the second method there is nothing like that
so now which method is faster in performance, does the number of function calls in first method reduce the performance......?
thanks and regards,
sri ram.

Oracle has a separate SQL and PL/SQL engine. When you go from one to the other, this is called a 'context switch'.
A context switch is a costly operation.
In variant 1 you seem to embed the function you wrote in SQL.
This means
- for each record you will have a context switch
- in the function you call SQL: another context switch
- The SQL in your function will show as recursive SQL and will never get merged into the main query.
So actually variant 1 is a textbook example of 'slow by slow programming' or 'How to write unscalable applications'
Remember Tom Kytes mantra:
When you can do it in SQL, you should do it in SQL
When you can't do it in SQL, you should do it in PL/SQL
When you can't do it in PL/SQL, you should do it in Java (in the database)
When you can't do it in Java, you should do it in any language outside the database.
Sybrand Bakker
Senior Oracle DBA

NAM Report Analyze -- Wan Optimization -- Application Performance Analysis Report

Hello,
Transaction Time (Client Exprerience) report does not show optimized traffic. For this graph what we expect is to show side by side the transaction time of optimized and non optimized traffic side by side to give information about the baseline and optimized application performance.
May it be that the optimization should be disabled for some time for the Wan Sites, and after some time manually reenabled in order to compare these values? This should be difficult for large deployments since manually disabling and reenabling for performance analysis time taking task.
Best Regards,

Mohammed,
It is common, in many of the Cisco Express 8.5 environments we have looked at, for the Total Incoming Calls given on a Traffic Analysis report to be a higher number than an Application Report.
The Traffic Analysis Report counts every unique sessionID (unique call) that is inbound (contact type of 1). The Application Reports do a similar thing but qualify (filter) only the records that have an application assigned.
There are simply times where inbound calls have been directed to an "agent" without having an applicaiton assigned.
The best thing the reporting user can do is to run a query on his or her database such as:
select * from ContactCallDetail where contactType=1and startDateTime > '2012-11-16 10:00:00' and startDateTime < '2012-11-16 11:00:00';
Usually when an application is not assigned to the record in the ContactCallDetail table it is because the destination type is equal to 1, which is an 'Agent' instead of a 'route point'.
So if you modify your select statement to filter by destinationType, you can quickly find the records that don't have the application assigned.
Example: select * from ContactCallDetail where contactType = 1 and destinationType = 1 and startDateTime > '2012-11-16 10:00:00 and startDateTime < 2012-11-16 11:00:00';
When you look at these records, you will see the agent that took the call from the destinationID field. The number in that field should match up with the field called 'resourceID' in a table called 'resouce';
Example: select * from resource where resouce=6011; where 6011 was the number you found in the destinationID field.
If there is still confusion about the source of the call - then talk to that agent and find out what is was.
Good Luck and let me know if you need further help.
Ron Reif
[email protected]

OBIEE 11g Application Performance Monitoring

Are there any best practices or recommendations for Application Performance Monitoring for OBIEE 11g?
I see that from Fusion MiddleWare Control, we can navigate to Business Intelligence -> CoreApplication (right click) -> Monitoring -> Performance
And from here we have a huge amount of performance metrics to choose from.
What I'm interested in finding out is:
1). Out of these hundreds of metrics, are there a subset of "core" performance metrics that Oracle recommends monitoring that gives a good baseline for performance of the application?
2) What are the options for capturing/gathering/saving these metrics for historical and trend analysis?

Excellent post. Thanks. Is there some sort of trial of the tool you guys built that's described in the post you linked?
"In addition to the out-of-the-box options above, here at RittmanMead we have developed our own OBIEE monitoring tool.
DMS metrics are stored directly on disk or through a database, enabling both immediate and retrospective analysis. Custom dashboards enable the display of both OBIEE and OS data side-by-side for ease of analysis. Integration with third-party tools is also an option."

DO's & DONT's for a better application performance

I am very new to tuning. What are the sql & pl/sql do's and dont's and best sql & pl/sql practices for a better application performance?
I was told by somebody to check explain plan for tuning db & sql statements. But i am bit confused by seeing the explain and don't know from where i have start and what i have take into consideration from explain plan to improve application preformance?
Can somebody help me in this regard?

> I am very new to tuning.
Good. Not having to tune means that the code is performing as it should.
And this is THE MOST IMPORTANT CONSIDERATION.
Do not write and design code that needs to be tuned and twisted and modded in order to perform. Write and design code with the core requirements of that code being:
- performance
- scalability
- flexibility
- maintainability
Good design = good code = good performance
No amount of tuning can turn bad code and a bad design into a well performing and scalable system. Which is why I dislike the concept of "tuning code". Start with a good and solid design. In other words, tune that before the code.
> What are the sql & pl/sql do's and dont's and best sql & pl/sql practices
for a better application performance?
The 1st basic rule is actually quite simple.
The "data layer" rule: Maximize SQL. Minimize PL/SQL.
Do not do in PL/SQL what SQL is very capable of doing. Fact. No other language is faster and better than SQL at crunching data in the Oracle RDBMS.
The 2nd basic rule is a logical extension of the 1st.
The "application layer" rule. Maximize PL/SQL. Minimize Java/C++/C#/Delphi/VB/etc.
Do not do in a 3rd party language what PL/SQL is extremely capable of doing. Fact. No 3rd party language is as capable as crunching data in the Oracle RDBMS as PL/SQL.
Stick to these two fundamental rules and your Oracle application will have a solid foundation for being a well performing and very scalable solution.

How to improve database and application performance

Hi,
Any body please help me out that how can we improve the database and Application performance.
Regards,
Bhatia

bhatia wrote:
Hi,
Any body please help me out that how can we improve the database and Application performance.
Regards,
Bhatiathere is no simple answer. There is no DATABASE_FAST=TRUE initialization parameter. There are a myriad of reasons why an application is performing poorly. It could be that the application (code and data relationships) is poorly designed. It could be that individual SQL statements are poorly written. It could be that you don't have enough cpu/memory/disk bandwidth/network bandwidth.
You need to determine the root cause of poor performance and address it. If you application is poorly designed, you can tune the database until the cows come home and it won't make any difference. If you are trying to run 100k updates per second against a database hosted on hardware that only meets minimal requirements to install Oracle ... well, hopefully you get the picture.
First, go to tahiti.oracle.com. Drill down to your selected Oracle product and version. There you will find the complete doc library. Find the Performance Tuning Guide
Second, go to amazon.com and browse titles by Tom Kyte and Cary Milsap. I particularly recommend "Effective Oracle by Design" and "Optimizing Oracle Performance", though I see a lot of new titles that look promising (I think I'll be doing some buying!)

How to get the Application perform actions when exits?

How to get the Application perform actions when user clicks on the "X" icon in the top right hand corner?
OR
If i placed an Exit Button.... when actions that i need to use to allow my application to perform a certain action when it exits?
Thanks

frame.addWindowListener(new WindowAdapter() {
public void windowClosing(WindowEvent e) {
// do your stuff here
});The WindowListener and WindowEvent can be found in java.awt.event package
//David

Oracle applications performance

Does anyone have any good best practices regarding how to increase
the performance of oracle applications ?
Found this resource
http://www.princetonsoftech.com/html_email/garfield/registerwebinar.asp?id=ggoracleapps
but other than this doesn anyone have other info/links?
Thanks,
Joe

Hi, buddy:
you have to be aware that oracle application performance depends on many points, including: oracle database(of course), network, OS(I/O,memory), and last but not least client SQL. while SQL problem may stand up to 50-70% of all.
performance tunning is a open-ended process, it won't turn out to be satisfactory overnight.
good luck!
eilison

Does native compiling improve application performance ?

Hi,
I've got familiar with Excelsior Native Compiler since a week ago. They claim on their website that compiling Java classes directly to native code (machine code) will improve the performance of the application. However, JAlbum (http://jalbum.net) says that its JAR files of the application run "basically at the same speed" compared to the native compiled one for windows.
Does really compiling Java classes to native code improve performance? I'm not talking about the startup speed, but the whole application performance.
Thanks...

Youve really been spreading that article aroundthis
morning, George, lol.
It was actually a really good read this morning.
Thanks for posting it.I know! just realized that's about the 3rd or 4th
time I've posted it today. just a coincidence, thoughcan't promote that book enough though. Should be required reading (just like Effective Java).

MDX formulae design

Hi,
I have two measures A and B. i need to calculate another measure C. which is ratio of A and B (A/B).
B is constant for all the combination
My problem is: if we apply more than one member filter from one of the dimension ( Web Analyzer):
I need a formulae like below:
(A for member1+A for member2+A for memebr3+.....)/B
Right now its working like below:
(A for member1+A for member2+A for memebr3+.....)/(B for member1+B for member2+B for memebr3+.....)
As B is a constant, i am loosing the definition of ratio.
Any idea how to implement will be of great help.
TIA

Hi Glenn,
I did that like
Added one measure:
"No Of Filters"
MDX formulae:
IIF(Count(Intersect({[DIMx].CurrentMember}, {Descendants([Gen2 Memeber of DIMx]),Descendants([Another Gen2 Memeber of DIMx])}))=0,0,1)
now what is happening is if i add multiple members as filters from another dimension the values multiplied.
As long as single members i have no problem. once i put multiple members in filter section that count is not correct. its getting multiplied.
Suppose say : 2 members from DIM1, 2 memebrs from DIM2 and 2 members from DIMx then count is becoming as 2*2*2=8;
and i need actually count of number of filters from Dimx that is only 2. Here again i am breaking the defintion of ratio.
Is thers anyway to count how many filters from one dimension?
Again thanks Glenn for oyur input.

J2EE application performance bottlenecks

For anyone interested in learning how to get resolve J2EE application performance bottlenecks, found a great resource:
http://www.cyanea.com/email/throttle_form2.html
registering with them can have you win 1 of 3 iPod mini's

I agree with yawmark's response #1 in one of your evil spams http://forum.java.sun.com/thread.jsp?thread=514026&forum=54&message=2446641

ANN: Learn How to Enhance J2EE Application Performance with ESI

OTN's newly enhanced Virtual Shopping Mall application illustrates how ESI technology can improve J2EE 1.3 application performance and availability.
http://otn.oracle.com/sample_code/products/ias/web_cache/index.html
Cheers,
-Srikanth

This solution does not seem to work if it is CDSSO.

Oracle applications performance-1

Does anyone have any good best practices regarding how to increase
the performance of oracle applications ?
Found this resource
http://www.princetonsoftech.com/html_email/garfield/registerwebinar.asp?id=ggoracleapps
but other than this doesn anyone have other info/links?
Thanks,
Joe

Hi, buddy:
you have to be aware that oracle application performance depends on many points, including: oracle database(of course), network, OS(I/O,memory), and last but not least client SQL. while SQL problem may stand up to 50-70% of all.
performance tunning is a open-ended process, it won't turn out to be satisfactory overnight.
good luck!
eilison

MDX formulae reduces application performance

Similar Messages

Maybe you are looking for