RSRT statistics data - high query runtime

Hi
I have a query on a multiprovider that has a few fields with exception aggregation upto 2 levels - doc number and item. I executed the query in RSRT with "Display Statistics data" and "Do not use cache" settings from the debug options. When i go to the Aggregation Layer tab of the stats screen i see that the basic infoprovider and the aggregates are listed twice or thrice. The infoproviders have a huge volume of data and the number of records read and transported for the duplicate entries shown are the same.
Initially i thought it is because of the 2 levels of exception aggregation but then i ran another query on the same multiprovider with similar exception aggregation and found that there are no duplicate entries and the basic inproviders are read only once.
the query is running very slow and i need to tune up the performance.
I would like to know why is the infoprovider being read twice for this query alone??
Regards,
Sujai

ok! Then it could be because of some other reasons why it is reading twice. This won't be happening because of nested Exception aggregarion.
Try to make the query same one by one and check what is making the query to read the infoprovider twice.

Similar Messages

Detailed Query Runtime Statistics - Substeps of Data manager Event

Dear All,
I am using the report for 0TCT_MC02_Q0200 to analyse the Detailed Query Runtime Statistics.
The report shows that the Event 9000 (Data manager Event) took a time of 109 seconds.
I would like to know as to what does the 9000 Data Manager Event that takes 109 seconds do for me.
What is it doing for 109 seconds ????
Can I have more details about the substeps of Data manager Event and time taken by them ??? Any Table or a report would suffice my requriement.
Thanks
Regards
Shalabh

Hi Jain,
Execute the query in RSRT or in RRMX. To Check how much time it is taking to execute this querry?? To get this information, Goto SE11> Give table name RSDDSTAT_DM in BI7.0 or RSDDSTAT in BW3.x> Display -> Contents-> Give from date and to date values as today, user name as Ur user name, and give the query name--> execute.
Now u'll get a list with fields like Object name(Report name), Time read, Infoprovider name(Multiprovider), Partprovider name (Cube), Aggregate name... etc. If the time read is less than 100,000,000 (100 sec) is acceptable. If the time read is more than 100 sec then it is recommended to create Aggregates for that query to increase performance.
Here in this table u'll find detailed Query Runtime Statistics.
Hope thsi helps you..
Regards,
Ramki.

Query Runtime Statistics - Delta Load Errors

Good Morning,
We recently upgraded our BI system to 7.0. One of the consultants who was here activated some standard business content, as the title of my subject shows, Query Runtime Statistics. The Infocube that is being loaded in the process chain he activated is: 0TCT_C01.   This load utilizes 4 individual delta loads, coming from Datasources: 0TCT_DSA1, 0TCT_DSO1, 0TCT_DSO2, and 0TCT_DSO3.
The problem we're having is that at least 2 of these frequently fail ( 0TCT_DSO2 and 0TCT_DSO3 ).
The failure messages we get is that some of the data coming in for object 0TCTSTATOBJ (Query Runtime Object) is always bad.   Some example values of the data that it is trying to load include:
FILTER(1)
TABELLE(1)
GENERISCHER NAVIGATIONSBLOCK(1
I'm not seeing anything in particular that would indicate these as being invalid values. The ()'s I wouldn't think would be causing it, given that those are in our system as acceptable characters (checked RSKC to verify) and there's nothing else that looks out of place (aside from the values coming in as German). Has anyone else come across an error like this for this particular statistics load? Manually changing these erroneous in the PSA is not an option given the frequency of the failures.
Any help would be greatly appreciated. If I need to clarify or relay any additional information, please let me know as well. Thanks

Sita,
I greatly appreciate your response. However, I probably should have provided some additional technical information regarding the issue.
The values I posted actually are from the PSA. These values that are given are telling the person what Java object the report being used, has had an issue (may not just be reporting issues, but general statistics for the use of these java objects, not sure yet).
I am aware that when BI gets an invalid character, it replaces it (in the display screen) with a #. Unfortunately, we do have # in our permitted characters and the reason why that doesn't work to qualify what was entered is because while it shows up in your display screen as a #, in the background it still sees it as the invalid char.   Subsequently, you won't always be able to see what that invalid char is. We've encountered this many times over the years with our Sales Orders when a Rep (or someone within the company) enters invalid characters (typically they do a copy and paste from Excel using a different Font, or are using arabic or other foreign characters).
Given that the values we're recieving are in German, my only conclusion is that the words themselves contain german characters often used in certain letters. For example, the letter U can sometimes have what they call an umlout (spelling may be off on that). An umlout almost looks like a quotation mark that is placed directly above the letter itself.   Another common letter in German is the ß symbol.
We may be able to bypass our problem by entering in these german characters, but I guess I was looking for more than just a "hack" solution and was curious if anyone else out there has had a similar issue on their system and if they were able to come up with a solution other than trying to guess and add in "special characters" that are not normally used in the english language.
Again, my appologies for not being more descriptive and thank you again for replying.

Error in BI Administration Cockpit for "Query Runtime Statistics"

Dear all,
after implementing the BI Administration Cockpit in BW and the EP I noticed, that whenever I try to modify my object selections under the node "Query Runtime Statistics", I get an error saying
"The Entered Value Is Not in the List of Valid Values"
This occurs under the following nodes:
   BI Application Statistics -> BI Application Selection -> Web Templates / Workbooks
   BI Application Object Statistics -> BI Application Object Selection -> Queries / Query Views / Planning Functions ....
   InfoProvider Statistics -> InfoProvider Type Selection
Any ideas how to fix this issue???
Thanks,
Andreas

Andreas,
Thank you for your response. I applied the note 1463442 to our Development system. It fixed the error, that is , now I don't get "entered value is not in list of valid value". But when I unselect objects and click on "SUBMIT" button, nothing happens. It should redraw the report based on the selection.
Is this working for you ? If so, please advise us.
Thanks again,
Ravi

Long Query Runtime/Web-template Loading time

Hi,
We are having a very critical performance issue, i.e. long query runtime, which is certainly not acceptable by client as well.
Background Information
We are using web application designer (WAD) 2004s release to design front end of our reports built in BI 7.0 system.
Problem Area
Loading of web template on browser
Problem Analysis
Query taking so long time to run, whenever we load it through portal or even directly through web application designer. Current runtime for query is more than a min. And I have noticed that 95% of runtime is taken for loading variable screen. FYI if I run query through Query Designer or BEx Analyzer, it takes 3-5 seconds to execute.
We have taken all the statistics and everything proves that query is not taking any time to execute but its the loading time which creates bottle neck.
Possible Cause
Web template holding 11 data providers, 5 of which are based on queries and rest are on query views. These data providers load into memory in parallel which could cause delay.
These data providers expose detailed variable screens. Out of 21 input fields, exposed by web template, 8 fields are based on hierarchy node variables and 1 on hierarchy variable. And to my knowledge each time hierarchy/hierarchy node variable loads complete hierarchy into memory whenever they are called (in other words, its not performance efficient to use hierarchies).
I request you to please consider this as matter of high priority and provide me with suggestions to remove bottle necks and make the application performance efficient. Please let me know, if you need any further information.
Thanks.
Shabbar

I would recommend you see how long the query execution actually takes without running from the web template. If actually the individual query takes long time then you need to do some performance improvement on back-end side (aggregates, indexing,... and so on).
But the performance issue is only with web templates, then you need to find some notes on it, because, I remember we had to apply some notes in relations of browser taking too long time to load the selection screen in web reports.
After exhausting all the option, then I will implement precalculating the query result before hand using broadcaster.
thanks.
Wond

Query Runtime evaluation against MultiProviders

I am testing dividing a cube into 2 logical partitions by fiscal year. I am placing a multiprovider on top of the logical partitions and have copied three queries I wish to test the runtime affect. This is a proof-of-concept idea I need to prove before suggesting an architecture change.
When executing the queries against the original cube, I can see the runtime results in ST03N. When I execute the queries against the MultiProvider, I do not see the results. ST03N only shows queries that were executed against a cube... not a MultiProvider.
I did try using RSRTRACE, but I am not confident the CMD_PROCESS time is correct. I have also looked into the standard delivered Statistical Reports, but didn't see any that would provide me with the runtimes of the MultiProvider queries. Perhaps I just missed the correct one.
Any recommendations?

Michael,
You could use RSRT which has a boatload of options on how you want the query to run when you first select Execute and Debug button (e.g. Do not use Cache, Do Not use BIA, etc.). One of my favorite options is Display Statistics Data which breaks down the query execution time into all of its myriad of components. You need to Green arrow back after you get the results to get to the Statistics screen.
Hope this helps.
Peace,
Dan

We are interested to know how/where to check query runtimes for any Query in SAP BW?

We are interested to know how/where to check query runtimes for any Query in SAP BW?
Is there any table or program to get the query run time details per query for a particular day.

Hello Sravan,
Bex statics tables would be 'RSDDSTATHEADER','RSDDSTATINFO' & 'RSDDSTATEVDATA'. all these can found in one view 'RSDDSTAT_OLAP'.
above can give historical statistic.
For current Query statics : T-code RSRT, and select debug mode select display statistics and then execute.
Once you get the output press F3(one step back), there you can see statics of the query for that particular execution.
Thanks,
Mallikarjuna

Semaphore statistics very high please help

Dear All,
In Semaphore statistics very high .please guide .how to solved it.
we are using windows 2003 and Ecc5
Semaphore statistics Date 05.09.2008 Time 13.27.22 Monitoring Options:
Wait time entering the critical path
Residence period in the critical path
Key No. <100us <1ms <10ms <100ms <1s <10s >10s Time/ms Protected area
2 329 327 2 0 0 0 0 0 0 Message Queue (Writing Dialog Requests)
2 329 329 0 0 0 0 0 0 0
3 2,040 2,040 0 0 0 0 0 0 0 Message Queue (Read)
3 2,040 2,040 0 0 0 0 0 0 0
4 11 11 0 0 0 0 0 0 0 Terminal communication
4 11 11 0 0 0 0 0 0 0
5 15 15 0 0 0 0 0 0 0 Work process communication
5 15 15 0 0 0 0 0 0 0
6 750 750 0 0 0 0 0 0 0 Roll administration
6 750 750 0 0 0 0 0 0 0
7 1,660 1,659 1 0 0 0 0 0 0 Paging administration
7 1,660 1,189 471 0 0 0 0 0 0
12 180 180 0 0 0 0 0 0 0 Character code conversion
12 180 179 1 0 0 0 0 0 0
13 200 200 0 0 0 0 0 0 0 Update synchronization
13 200 200 0 0 0 0 0 0 0
14 2,161 2,125 1 11 18 6 0 0 1098 Presentation buffer
14 2,161 1,290 829 9 29 4 0 0 2061
16 1,028 1,026 2 0 0 0 0 0 0 Table buffer key
16 1,028 983 34 11 0 0 0 0 0
17 1 1 0 0 0 0 0 0 0 Buffer synchronization
17 1 0 0 1 0 0 0 0 0
23 1,578 1,578 0 0 0 0 0 0 0 Message queue (Writing All Other Requests)
23 1,578 1,576 2 0 0 0 0 0 0
24 11 11 0 0 0 0 0 0 0 LRU table buffer
24 11 11 0 0 0 0 0 0 0
26 1,042 1,042 0 0 0 0 0 0 0 Enqueue table
26 1,042 1,034 6 2 0 0 0 0 0
31 255 255 0 0 0 0 0 0 0 Spool administration
31 255 254 1 0 0 0 0 0 0
33 16 16 0 0 0 0 0 0 0 Extended segments management
33 16 16 0 0 0 0 0 0 0
34 2 2 0 0 0 0 0 0 0 Message buffer
34 2 2 0 0 0 0 0 0 0
38 1,809 1,809 0 0 0 0 0 0 0 CCMS Monitoring
38 1,809 1,724 85 0 0 0 0 0 0
39 1,736 1,736 0 0 0 0 0 0 0 Extended Global Memory
39 1,736 1,736 0 0 0 0 0 0 0
42 430 430 0 0 0 0 0 0 0 Statistics Buffer
42 430 430 0 0 0 0 0 0 0
43 141 141 0 0 0 0 0 0 0 Spool Cache
43 141 138 2 1 0 0 0 0 0
46 1 1 0 0 0 0 0 0 0 Profiles
46 1 1 0 0 0 0 0 0 0
48 417 417 0 0 0 0 0 0 0 EnqID
48 417 417 0 0 0 0 0 0 0
57 1,102 1,102 0 0 0 0 0 0 0 Runtime Monitor
57 1,102 1,101 1 0 0 0 0 0 0
Thanks,
Kumar
Edited by: Kumar on Sep 5, 2008 11:17 AM

Hello Bhaskar
The above log is of Semaphore in St02 deatils.
we are using windwos 2003 ,Ecc5 and oracle 9i.
I am geting Sem in Sm50. and by obervation i am getting number 14, 38 and 7 manny times . form note 33873
7: PAGING-Semaphore (paging administration)
14: PRES_BUF-Semaphore (presentation buffer)
38: SEM_CCMS_AS_MONI_KEY (CCMS monitoring for appl.s)
But i need sugging how to solve the problem .
the other log are os ST06 and St02.
I gont have much free space.
The problem is system is working very slow and getting hang.
In st06 the Lan high.
Lan
(sum)
Packets in/s 4,444 Errors in/s 0
Packets out/s 4,170 Errors out/s 3
Collisions 0
Memory Physical mem free Kb 1,788,888 in the system so i cant increase
IN ST02 Export/import swap is high than 10,000.
Export/import 89.72 40,000 7,528 25.77 30,000 18,855 62.85 95,972 Exp./Imp. SHM 97.89 4,096 3,238 95.94 2,000 1,999 99.95 0 0
Thanks,
kumar

Dynamic Data Binding at runtime

For future applications, flexibility will be an important feature. One of the biggest drawbacks of forms is/was, on my opinion, not being able to create new items at runtime.
JClient will easily overcome this restriction. But how about data binding ?
Imagine the following situation: We've got a VO "order" with the "custname" incorporated as lookup data. At runtime, the user decides that she needs to see the "custloc" as well (put an additional JTextField into some whitespace on his panel). As I understand it is possible to build a view link to tie an instance of the VO "customer" to the "order" (I'd prefer that over adapting the query statement of "order" at runtime). But how can I create the new iterator binding and control binding needed to populate the dynamically created attributes with the data from "customer"? And, further on, how will I be able to make the changed configuration of my panel persistent, let's say into the DB ? Or does the data binding concept of 10g just not cover this requirement?

Hi pascal
It is not possible to send you email at your address. All come back with a permanent error
Your document: test
was not delivered to: <[email protected]>
because: Error transferring to smtp.solnet.CH; SMTP Protocol Returned a Permanent Error 550 Service unavailable; Client host [81.62.5.7] blocked using dul.dnsbl.sorbs.net; Dynamic IP Address See: http://www.dnsbl.sorbs.net/cgi-bin/lookup?IP=81.62.5.7

OWL with Default Set containting Current Date as query parameter

Hello Everybody,
I have a custom BO which mainly consists of a name and a time period (start and end date).
Now I'd like to create a Default set like
"Give me all objects with today between start and end date."
I already tried the following:
Filling a field using UI-Scripting:
$data.UIState.Today = $controller.DateTimeUtils.DateNow;
Copy the field into the search parameter
executing the query with this search parameter.
However, I using this approach, I only managed to search for objects matching exactly todays date, but not using greater/lower comparators.
Furthermore I also searched for a field transformation giving me the current date but did not succeed.
Does anyone have any ideas how to achieve this or how to get one of the above ideas to work?
Best regards,
Ludger

Hello Ludger,
That`s what I expected: The fields do not define the interval.
Therefore you need to define it by yourself via this "less or equal".
What you can do in the UI Designer at the query definition is to provide an interval
For the field "CurrentDate" provide
- SelectOption = "between"
- Low Value = your start date
- High Value = your end date
- Select Option = "include"
HTH,
Horst

Dynamic calculation of values at query runtime

Hello, Everybody,
I have situation in which I have calculate the value for characteristic at query runtime.
Query is based on Infoset SEMPLDATA. In Infoset I have fields u201CEmployee IDu201D, u201CBirth dateu201D, u201CAgeu201D dedicated as virtual variable, u201CNumber of Employeesu201D.
Then I execute query I enter a value for variable u201CDateu201D. After that query should calculate age for employee: for example ( variable u201CDateu201D u2013 u201CBirth dateu201D )/365. And query result is listed like this: u201CAgeu201D, u201CNumber of Employeesu201D.u201D Ageu201D is dynamic it depends on variable u201CDateu201D.
Iu2019m doubting will virtual characteristic help me. Could you give your suggestions?
Thanks in advance.
Best Regards,
Arunas Stonys

Hey. You can do this with a virtual characteristic. I'm not fond of them because of performance, but it should work. I am also assuming you have date of birth as an available field to do the calculation.
So, you could implement code in the variable user exit. After "Date" has been entered and the exit code is invoked you can pass that date to memory. Something like export "date" to memory ID xxxxx. I would create the memory ID object based on user ID running the query and date, or something like that. Then when you are processing in the virtual characteristic user exit, just do an import from memory, referencing the same memory ID as when you exported. Then you should be able to just do the date of birth - "date."
I have done this in the past to be able to retain user selection information for use in the virtual object calculations.
Just thought I would add a bit more here. You could also use formula variables based on your "Date" variable and "Date of birth" and do the calculation in a formula. Not completely sure about your situation, but that's another idea.
Hope this helps.
Thanks
Edited by: Siboda_Eric on May 12, 2010 4:21 PM

Unit of Measure conversion in query runtime

Hi all,
I am attemoting to utilise the ability to convert key figure quantities in BI7, at query runtime by using a user entry variable and a conversion type.
Does anybody know of any decent documentation as to how to configure this? The SAP help documentation seems both confusing and inaccurate.
My data records contain key figures with unit 0BASE_UOM and conversion between units for each Material code are held in 0MAT_UNIT.
Thanks in advance

Hi
My feeling is when you do conversion on aggregated level(multiple products), if any of the product has UOM missing then you coversion fails on aggregted level.
It looks strange that UOM conversion looks fine when you do conversion on each product.
I suggest you to check whether the UOMs (that you want to convert)are updated correctly on each product on their repective location in the product master under UOM tab.
regards
Vijaya.

Short dump while loading statistics data - 'CX_RSR_X_MESSAGE'

Hi all,
When I am loading statistics data with Init selection, system gives short dump error. Below is the message-
The exception 'CX_RSR_X_MESSAGE' was raised, but it was not caught anywhere
along
the call hierarchy.
Since exceptions represent error situations and this error was not
adequately responded to, the running ABAP program 'SAPLRRMS' has to be
terminated.
If anyone know the answer please share with me.
Regards,

Hi Preet,
Please find the details of short dump from ST22-
Error analysis
    An exception occurred which is explained in detail below.
    The exception, which is assigned to class 'CX_RSR_X_MESSAGE', was not caught
     and therefore caused a runtime error.
The reason for the exception is:
No text available for this exception
to correct the error
If the error occures in a non-modified SAP program, you may be able to
find an interim solution in an SAP Note.
If you have access to SAP Notes, carry out a search with the following
keywords:
"UNCAUGHT_EXCEPTION" "CX_RSR_X_MESSAGE"
"SAPLRRMS" or "LRRMSU13"
"RRMS_X_MESSAGE"
Regards,

Last changed date of query showing old date

Hi All,
We have made a change to a query (description change) and saved. When we open the Analyzer we still see the old description and also the last changed date shows an old date. Not the latest date the query was last changed on. We tried generating the query in RSRT, but no luck.
Can someone let us know what we can do?
Thanks!

Hi Kams,
We are using the Analyzer, log out back in, etc...
As mentioned we did a regenerate by RSRT 'generate'
any other ideas?

QUERY RUNTIME

I have a query which is based on the ODS.The query is taking a lot of time to run.What parameters should i check and how should i make sure that the query will run quickly.Is it something to do with indexes in the ODS.

I saw this information which might be useful
As stated by several people, primary indices are supplied with the ODS objects. With regard to to query performance on ODS objects, the best approach is to trace (st05) the execution of a poorly performance access. When the trace is read, normally the system will spend a long time on one access (it can vary, but this is most common). A secondary index can then be created on the selective fields in the long running WHERE clause. For example, I recently improved the performance of a query which was hanging for 32 seconds to 2 seconds. The SQL looked similar to this:-
SELECT
FROM
"/BI0/APU_O3200" T_00 , "/BI0/SCURRENCY" T_01 , "/BI0/SCMMT_ITEM" T_02 , "/BI0/SFUNDS_CTR" T_03 ,
"/BI0/SPU_MEASURE" T_04 , "/BI0/SAC_DOC_TYP" T_05 , "/BI0/SDATE" T_06 , "/BI0/SAC_DOC_NO" T_07 ,
"/BI0/SPSTNG_SEQ" T_08 , "/BIC/SZTXTLG" T_09
WHERE
( T_00 . "FM_CURR" = T_01 . "CURRENCY" ) AND ( T_00 . "CMMT_ITEM" = T_02 . "CMMT_ITEM" AND T_00 .
"FM_AREA" = T_02 . "FM_AREA" ) AND ( T_00 . "FUNDS_CTR" = T_03 . "FUNDS_CTR" AND T_00 . "FM_AREA"
= T_03 . "FM_AREA" ) AND ( T_00 . "PU_MEASURE" = T_04 . "PU_MEASURE" AND T_00 . "FM_AREA" = T_04
. "FM_AREA" ) AND ( T_00 . "AC_DOC_TYP" = T_05 . "AC_DOC_TYP" ) AND ( T_00 . "PSTNG_DATE" = T_06
. "DATE0" ) AND ( T_00 . "ACDOC_NO_F" = T_07 . "AC_DOC_NO" ) AND ( T_00 . "PSTNG_SEQ" = T_08 .
"PSTNG_SEQ" ) AND ( T_00 . "/BIC/ZTXTLG" = T_09 . "/BIC/ZTXTLG" ) AND T_00 . "FISCPER3" = '004'
AND T_00 . "FISCYEAR" = '2005' AND ( T_00 . "FM_ACTDETL" = '000' OR T_00 . "FM_ACTDETL" <> '000' )
AND NOT ( T_00 . "FM_ACTDETL" = '010' ) AND T_00 . "FM_AREA" = 'F1' AND T_03 . "SID" = 2684 AND
T_04 . "SID" = 1281
GROUP BY
T_01 . "SID" , T_02 . "SID" , T_03 . "SID" , T_04 . "SID" , T_05 . "SID" , T_06 . "SID" ,
T_07 . "SID" , T_08 . "SID" , T_09 . "SID" , T_00 . "FISCPER3" , T_00 . "FM_VTYPE"
6 TABLE ACCESS FULL /BI0/APU_O3200
( Estim. Costs = 20,821 , Estim. #Rows = 1 )
Don't be put off, the only imporatnt parts are the ODS table name (in the FROM clause) - "/BI0/APU_O3200" in this case and the end of the WHERE clause. The start of the WHERE clause specifies the join conditions (which we NOT interested in) and the end gives us the (hopefully)selective fields which access the ODS object ("/BI0/APU_O3200"). Here the selective fields are:-
FISCYEAR
FM_ACTDETL
FM_AREA
A secondary index should be created on these fields to improve performance and avoid the full table scan seen above. If query runtime is still unacceptably high, then more selective fields will have to be added, which of course is also a functional issue.

RSRT statistics data - high query runtime

Similar Messages

Maybe you are looking for