Infocubes?ODS?Multiprovider?Infoset?-Performance?????

Hello Experts,
      Infocubes?ODS?Multiprovider?Infoset?
In our design stage we had no problem creating ODS and Cube's but as our requirements kept changing we had to create Infosets rather than Multicubes as we were not having common fields between Cube and ODS. Secondly we couldn't have cubes as Infosets are built between ODS and Infoobjects.
      Can someone throw some light on the performance issues based on our current design were we have ODS and Infosets built. What are the pros and cons between
1) InfoCubes and ODS
2) MultiCubes and Infosets
      Any kind of thoughts would be of great help in our analysis.
Thanks,
Priya
Guys anyone who faced difficulty in analyzing such issues please do post your views.

We has the same requirements ...and had to build a lot of info sets ....
Reporting on any of the Navigational Attributes is NOT possible on an Info Set.
Go to RSRT and look on what fields on which the query is doing a Sequential read...if required try to build indexes on that fields.
In some cases we cannot restrict the queries ...if you are queering on yearly reports ...running against InfoSet's is not at all recommended......
For better performance of your queries try to build secondary indexes and as far as possible try to avoid InfoSet ....BW 3.5 has ADP try to workout something on that if u are on 3.5.
For performance try to compress your request in your cubes on your  BP.
Let me know
Message was edited by: Suresh Kaudi
Message was edited by: Suresh Kaudi

Similar Messages

  • I need to load an InfoCube from an InfoSet - Is this possible in 2004s?

    Greetings,
    <b>This question looks long, but it's really not.  I'm just trying to give a lot of detail about the screens that I am seeing.  Anyway, here goes:</b>
    I am in 2004s.  I want to load an InfoCube from an InfoSet.  I know that an InfoSet is just a view and does not physically store data, but recent documentation published by SAP seems to indicate that this is
    possible in 2004s (There is a document called "Developer Guide - 2004s" that says the following:
    "InfoSets are an additional, easily manageable data source for generic data extraction.  They allow you to use logical databases of all SAP applications, table
    joins, and further datasets as data sources for BI".  ) 
    So... that led me to believe an InfoSet could load an InfoCube.  In other words, the InfoSet would be the source and the InfoCube would be the target.  I am in RSA1, and I right-click on the cube that I want to be the target.  I select the "Create Transformation" option from the context menu.  In the "Create Transformation" screen that pops up, I enter "InfoCube" in the target object type field and the name of the InfoCube in the target name field.  I enter "InfoSet" in the source object type field and the name of the InfoSet in the source name field. 
    I get a pop-up that says "Cannot generate proposal".  I click the green arrow, and then I get an error message
    that says "Transformation does not exist (see long text)". 
    When I click on the error message, I get the following information in the Performance Assistant:
    "Transformation does not exist (see long text)
    Message no. RSTRAN010
    Diagnosis
    The transformation specified by transformation ID , source  and target , does not exist.
    System Response
    The system terminates processing."
    Any help would be much appreciated!  It seems like it must be possible to use an InfoSet as a source in a data load, or else it would not be an option in the "Create Transformation" screen.  ????
    I've also tried right-clicking on the cube and selecting "Create Data Transfer Process".  I enter the same information on this screen: the types and names of the source and target.  Then I get a pop-up that says "Source and target have not been linked with a transformationDo you want to generate a default transformation?".  There are two buttons I can click: one labeled "Yes" and one labeled "Abbrechen".  I choose "Yes", because I have no idea what "Abbrechen" means.
    Then I get the same pop-up from before: "Cannot generate proposal".  I hit the green arrow, and I'm back in the
    "Create Data Transfer Process" screen.  So... I'm stuck in an infinite loop of screens that I can't get out of. 
    If someone could help, I would be grateful!  THANK YOU VERY MUCH!

    Thank you so much, Ram!  I REALLY APPRECIATE YOUR HELP!
    I am wondering if it possible to manually create a node in RSA6 for the InfoSet.  I tried it by doing the following:
    I went to RSA6 and clicked on the white icon for "Create Node".  I then entered the number "8", followed  by the name of the InfoSet.  (I entered an "8", because all our DataSources seem to start with an "8".  I'm not sure why.) 
    The DataSource now shows up, but when I double-click on it, nothing happens.  When I select it in the hierarchy and then click the icons for "Display" or "Change" or "Check", I just get an error message that says, "You have not selected any DataSources."
    Also, do you know where I can get my hands on the white paper? 
    Thanks again!
    Regards,
    Sarah-Jane

  • Infocube & ODS Difference

    Hi,
    What is the difference between Infocube & ODS ? and in which senario they can be used ?
    Regards
    Praveen

    Hi Praveen
    ods
    1 flat file formate structure
    2 two dimentional
    3 overwrite data functionality
    4 performance is less as compared to cube
    5 detailed form of data
    infocube
    1 star schema
    2 16 dimentional
    3 additive data functionality
    4 performance is better as compared to ods
    5 summerised form of data
    Assign points if it helps
    Thanks & Regards
    santosh

  • DELETION OF SPECIFIC COLUMN DATA IN INFOCUBE&ODS(not the datarecords)

    hai
    I loaded the data into ODS and InfoCube.Buti want to delete the 'year = 1900&1910" from the YEAR column of the Infocube and ODS.
    Is it possible to delete the specific column data from the InfoCube & ODS????
    If it is possible then how can i delete the particular specific data from the specific column of InfoCube and ODS.
    pls tell me
    i ll assing the points
    bye
    rizwan

    HAI Ronald
    I already loaded the data into the InfoCube and ODS. But i want to delete the 'year column=1900 , 1910' only (not the corresponding datarecords).
    IS it possible to delete that Year Column only from the loaded data of InfoCube .
    I think , it is possible by creating a new InfoCube and transferring my InfoCube data into newly created InfoCube and writing the start routine for Update Rules in between  my InfoCube & newly created InfoCube.
    So do you know the ABAP/4 code for that
    pls tell me
    i ll assing the points
    bye
    rizwan

  • Finding the infocube/ods space

    Hi,
    Using transaction ST14, I have got the Space used by top 30 Infocubes, ODS etc. on quality server. The same I tried to do on production server. But it gives the message
    "Field or table is initial" instead of the list of Infocubes or ODS.
    But actually everything on quality and production server is same. so i am not getting the proper solution.
    Can anybody tell me how can i solve this problem?

    you need to schedule the job in ST14 to get the first result for top 30 objects...but take care..as it will take a lot of time and resources and your other scheduled jobs will starve...
    Also you can get the size through DB02...object name and then add the size of all the tables pertaining to that object.
    Hope this helps.
    Regards
    Gajendra

  • Designing a InfoCube / ODS

    Hi
    I would appreciate, if you can provide a solution to the following questions.. 
    1) What are the main consideration while designing an Infocube / ODS ?
    2) Do I need to add 0RECORDMODE InfoObject manually, while designing a ODS. ? Please give me step by step document for creating a Infocube and ODS Objects.
    Please Email me at [email protected]
    Thanks
    Preeti

    Hi,
    Designing ODS/CUBE depends on u r reporting requirement.
    Sometimes there is no ODS for specific functional area. If u r going to design the ODS then ou need to know the the table from where u r datasource is extracting the data. Then on the basis of that u will have to design u r ODS. If u r using 3.5 then u need to add RECORDMODE manually but in BI7.0 it will added as a default value initial.
    Then u can design all the relevant dimension in the cube on the basis of logically related infoobject. U can use line item dimension if any infoobject having same data record as fact table.Here u can follow Kimbal's rule while designing the Aggregates.
    If u r using BI 7.0 then u may think on BI accelerator to design and maintaing the cube.
    Thanks,
    Debasish

  • Loading performance of the infocube & ODS ?

    Hi Experts,
    Do we need to turn off the aggregates on the infocubes before loading so that it will decrease the loading time or it doesn't matter at all, I mean if we have aggregates created on the infocube..Is that gonna effect in anyway to the loading of the cube ? Also please let me know few tips to increase the loading performance of a cube/ods. Some of them are
    1. delete index and create index after loading.
    2. run paralled processes.
    3. compression of the infocube , how does the compression of an infocube decrease the loading time ?
    Please throw some light on the loading performance of the cube/ods.
    Thanks,

    Hi Daniel,
    Aggregates will not affect the data loading. Aggregates are just the views similar to InfoCube.
    As you mentioned some performance tuning options while loading data:
    Compression is just like archiving the InfoCube data. Once compressed, data cannot be decompressed. So need to ensure the data is correct b4 Compressing. When you compress the data, you will have some free space available, which will improve data loading performance.
    Other than the above options:
    1.If you have routines written at the transformation level, just check whether it is tuned properly.
    2.PSA partition size: In transaction RSCUSTV6 the size of each PSA partition can be defined. This size defines the number of records that must be exceeded to create a new PSA partition. One request is contained in one partition, even if its size exceeds the user-defined PSA size; several packages can be stored within one partition.
    The PSA is partitioned to enable fast deletion (DDL statement DROP PARTITION). Packages are not deleted physically until all packages in the same partition can be deleted.
    3. Export Datasource:The Export DataSource (or Data Mart interface) enables the data population of InfoCubes and ODS Objects out of other InfoCubes.
    The read operations of the export DataSource are single threaded (i.e. sequential). Note that during the read operations u2013 dependent on the complexity of the source InfoCube u2013 the initial time before data is retrieved (i.e. parsing, reading, sorting) can be significant.
    The posting to a subsequent DataTarget can be parallelized by ROIDOCPRMS settings for the u201Cmyselfu201D system. But note that several DataTargets cannot be populated in parallel; there is only parallelism within one DataTarget.
    Hope it helps!!!
    Thanks,
    Lavanya.

  • Query on Multiprovider after removing infocube from multiprovider

    I have a set of queries built on a multiprovider that contains 3 infocubes.  2 of the infocubes contain the same data (one is a new version replacing the old version) All of the queries are filtered to select the correct multiprovider.  I removed the old cube from the multiprovider and everything works fine but when I run RSRT on any of the queries and check the performance tab, I still see a reference to the old cube that is no longer part of the multiprovider. 
    If I create a new query on the multiprovider, RSRT does not reference the old cube.  There must be some rteference in the query definition to the old cube, how can I get rid of it?

    in RSRT use execute+debug option
    and under MP select MP explain and also select SQL run.
    It will show the IC or ODS the MP is using and the rest of the infoproviders will be skipped .
    this will give u a better idea wether ur query is accessing old cube or not.
    apart from this if ur MP still reference the IC but no query is accessing. it........................it is fine..
    However u can raise a ticket to SAP

  • Infoset - Performance and Design Question .

    Hello BW Gurus,
    I have a question regarding the infoset.
    We have a backlog report designed as follows:
    sd_c03 - with around 125 chracteristics and kf's - daily load will have around 10 K records at a time.
    custome cube - say - zcust01 with around 50 characters and KF's.- daily full load aroun 15 K records.
    My question:
    1.We have infoset ontop of this 2 cubes for reporting. Can we use infoset for reporting .I used infoset for master data and DSO reporting. Do you guys see any performance issue with using the infoset instead of multiprovider. Is there any alternative instead of reporting from infoset.
    2.Also I executed the  SAP_INFOCUBE_DESIGNS program for the above cube and some dimesions are more then 25% like 58%, 75%,even 102%. So this has to be fixed for sure. Is it correct.if we don't change the design then what will be the consequences.
    Please advise. We are in the development and the objects are not yet moved to production yet.Again thanks for your help in advance.
    Senthil

    Hi......
    1.We have infoset ontop of this 2 cubes for reporting. Can we use infoset for reporting .I used infoset for master data and DSO reporting. Do you guys see any performance issue with using the infoset instead of multiprovider. Is there any alternative instead of reporting from infoset
    Multiproviders are a great tool that can be used freely to "combine" data without adding much overhead to the system. You may need to combine data of different applications (like Plan and Actual). Another good use is from a data modeling point of view...you can enable "partitioning" of the data by using separate cubes for separate years of data. This way the cube are more manageable, and with an InfoProvider on top the data can be combined for reports if reqd (instead of having one huge cube with all the data).
    Also using a multiprovider as an 'umbrella' infoprovider, you can easily make changes to underlying InfoProviders with minimum imapct to the queries.
    You should go for multi-provider when the requirement is to report on combined data coming from different infoproviders.
    Multi-providers combines data using union operation, that means all values of underlying infoproviders are taken into consideration. On ther other hand, Infoset works on the principle of join. Join forms the intersection and only the common data from underlying infoprovider are considerd..............In terms of performance infoset is better since it fetch less number of records.........
    Check this .........
    https://www.sdn.sap.com/irj/sdn/go/portal/prtroot/docs/library/uuid/2f5aa43f-0c01-0010-a990-9641d3d4eef7
    2.Also I executed the SAP_INFOCUBE_DESIGNS program for the above cube and some dimesions are more then 25% like 58%, 75%,even 102%. So this has to be fixed for sure. Is it correct.if we don't change the design then what will be the consequences
    If Dimensions are more it effects query performance.........its better to change the design........
    Hope this helps........
    Regards,
    Debjani.........

  • Copy Query Variants from InfoCube to MultiProvider

    Hi,
    Is it possible to copy query variants from an infocube to a multiprovider?
    Basically, we used RSZC to copy the queries from the infocube to the multiprovider, but it doesn't look like the variants were copied over. Is there an ABAP program or some other workaround that can expedite/facilitate in recreating the variants on the multiprovider queries?
    Your help is much appreciated.
    Thanks,
    KK
    Edited by: Kristin Kalalau on Jan 29, 2009 10:48 PM

    It is possible to copy but the only requirement is your target cube structure should match with the source cube structure.
    Thanks...
    Shambhu
    Edited by: Shambhu Kumar Gupta on Jan 30, 2009 12:01 PM

  • ODS Settings and performance

    Could any of the following ODS settings individually or in combination have an impact on performance? 
    BEx Reporting
    Set quality status to 'OK' automatically
    Activate ODS object data automatically
    My ODS will have a maximum of 250,000 records and is only an InfoProvider to 4 queries and is not used for staging.  Thanks

    Hi Niten!
    If you built (as you said) some query on your ODS, you need (as you already know) BEx reporting checkbox flagged.
    Clearly this have some impact when you activate every request since the systems have to work with SIDs too (in order to have the reporting capabilities and to let OLAP processor work).
    Anyway, in order to identify the relevance of this impact, this depends from many other factors (how much powerful is the system you have...); but if every single request (as you wrote) contains only a few records, don't worry !
    Set... and Activate... flags: no performance impact because are a simple automation process...anyway, you have to activate your ODS!!!
    Flag always the first one "Set..." (generally speacking), flag "Activate..." if you want to start automatically the activation and you don't use the process chain, other wise don't flag it and insert the specific process in a PC...
    Hope now it's clearer...
    Bye,
    ROberto

  • Help: Meta documents for ODS/Multiprovider not visible in BEX

    If I created documents in rsa1 for basic cubes, these documents show up fine in BEX query property.
    But this function does not work for ODS or Multiprovider. I created documents for ODS and Multiproviders and verify they exist after saving them. In BEX, open 'Query Properties' and click 'Documents', no document is availabe.
    Can anyone help? Strange thing is it works for basic cubes but not ODS or multiprovider.
    I went through RSA1 'Document Administration' and did not find anything unusal.
    Any help is appreciated.

    Hi Karthik,
    If you find ur self difficult in finding the "role folder" in Bex, Just go ahead and create roles in PFCG.
    PFCG, create roles, change authorizations, take template S_RS_RREPU .. i think u know all this.
    Go to S_COMP, assign IA, IC, Query ID properly.
    Its going to work in same fashion as you deploy the query to reole.
    please close it if u feel ans worthful.
    Regards,
    Gowd.

  • HOW DOES COMPRESSION OF AN INFOCUBE INCREASE THE LOAD PERFORMANCE?

    Hi all
    I see that Compression of infocube is one of the parameters to improve/increase load performance. If I am not wrong, Can some one please explain how compressing a cube improves the load performance?
    Thanks in advance
    Rishi

    Hi,
    I see that Compression of infocube is one of the parameters to improve/increase load performance. If I am not wrong, Can some one please explain how compressing a cube improves the load performance?
    As per my information Compression improves the Quaer performance not loading perforamnce
    when u do compression the same characterstics which are having the same values those records will be moved to E Fact table
    ex  Custno   Mat NO Qty   Value
          C101     M101     10      100
          c101     m101      20      200
    when u do comperession the recors will be compressed as below
    c101   m101  30 300
    when the query execution instead of reading two reocrds and compressing at the time of producing output at report level
    but when already compressed it fetchs one  record directly form E table like that the  query perforamnce will be improved.
    Not loading perforamnce.
    Thansk & regards,
    sathish

  • Compress Infocube Real Time problem - performance

    Hello All!
    I am having a problem with a cube and not able to compress the requests , repare indices , save data in this cube......please help!!
    If I start the jobs they did not finish and if i try to run and save a planning sequence i got an error (OR A-14400 inserted partition key does not map at any partition.....the cube is not partioned.
    I also tried to repare in RSRV and it get locked , never end.
    This is real time infocube.
    Let me know your commnents guys
    Thank you
    RDP

    See if the below threads help you in some way:
    ORA-14400: inserted partition key does not map to  any partition
    "ORA-14400: inserted partition key does not map to any partition"

  • Probelm in Deleting INFOCUBE/ODS?

    Hi ,
    I'm trying to delete the Bussiness content CUBE and ODS,
    Getting error when to delete Cube "Infoprovider "xyz" is not use Externally"
    ODS error " Infoprovider "xyz" set to INACTIVE.
    Please advice how to delete the CUBE/ODS in BI7.0 version.
    Thanks,
    Srini.

    Hi,
    Check if your ODS is in active state?
    Thats what I understand from the error message.
    thanks,
    JituK

  • BW QA User Profile

    I have been asked to test a user id in BW D , which will the same profile to be used later on in BW Q. I know I have to make sure that there is no change / delete / create authorization .
    What are the other things I need to make sure of for this profile.
    Thanks

    Hi Jimmy,
    Start with 'create' options first. If you able to create an object (InfoCube, ODS, MultiProvider, InfoObject, Infoset, hierarchy, InfoPackage, InfoSource, transfer rules, update rules, etc.), then the next step is to 'change' that object, or 'delete' that object.
    If you need the ability to execute an infopackage, then you also need transactions SM37, SM50, RSMON (monitor).
    Some more scenarios:
    Replicate DataSources
    Process chains (create, execute)
    InfoPackage groups (create/schedule/remove)
    Caution: Always use the create option first. You wouldn't want to delete something that is custom created.
    Always document your steps, it helps.
    I have about 300 tests which I used for testing power user and administrator roles.
    Best regards.

Maybe you are looking for