Skip to Content
author's profile photo Former Member
0
Former Member

What r the ways to reduce the load time of master data

Posted on Oct 14, 2005 at 08:04 PM 17 Views

Hi

The full load of the master data load IP is taking 18hr to 2 days loading on average 2516360 Records

Is their any option to reduce load time as its causing data base lock and impacting transports(other than data selection)

Thanks in advance

Anuj

Add a comment
10|10000 characters needed characters exceeded

Assigned Tags

Related questions

3 Answers

  • author's profile photo Former Member
    Former Member
    Posted on Oct 14, 2005 at 09:15 PM
    0

    Hallo Anuj,

    It depends on the master data.

    are master data time depending?

    AndyML

    Add a comment
    10|10000 characters needed characters exceeded

  • author's profile photo Former Member
    Former Member
    Posted on Oct 15, 2005 at 12:26 AM
    0

    I would first review the load in the Monitor to confirm where the time is spent. Maybe it is not an issue with loading the data, but the extraction from R3. There might be something on the R3 side that can be done, e.g. add index.

    I don't think number range buffering does you any good in this case, but it might help a little if you are adding more than several hundred new records daily.

    There is a Note on turning off <b><u>SAP buffering</u></b> for large <b><u>master data</u></b> tables. Don't have the Note handy, will try to dig it up but if you search on the highlighted terms you should get a hit.

    What DB are you using?

    Add a comment
    10|10000 characters needed characters exceeded

    • Former Member Former Member
      Oct 15, 2005 at 01:57 PM

      You should also be able to break up your load into multiple Infopackages with different selection criteria. Our 0EMPLOYEE MD load was taking a long time and we split into into several concurrent loads, e.g. employee 1 - 20000 in one load, 20001 - 40000 in another, etc. With Oracle, you should be able to run at least one load per CPU, maybe two, but that might be a trial and error process to find the right number. Might need to have the DBA increase the INITRANS for the table to handle a higher number of concurrent upates.

      Are there update rules that perform any table lookups during the load to enhance the MD? If so, they could be adding a lot of time - might be able to do something with them, e.g. loading the table to an internal table in a start rtn, and then reading the internal table in the update rule.

    Comment on This Answer
  • author's profile photo Former Member
    Former Member
    Posted on Oct 15, 2005 at 03:22 PM
    0

    You will have to do some research. What MD extractor are you talking about?

    Test on R/3 system: First try to extract a considerable amount of records via transaction RSA3 in R/3 (10.000 or 100.000). Keep the time that it takes.

    Extract the same into BW but only into PSA. Again measure the time.

    Load data from PSA into datatarget and see how long this takes. You should now have a picture on where the performance problems are located.

    Is the performance also bad for small loads or is there a boundary before which performance is ok. (in other words, is loading 200.000 records ten times longer then loading 20.000 records?)

    Suspect the following in R/3:

    - datasource enhancements in R/3. A redesign may improve your extraction in a big way.

    - missing indexes. If you are extracting data from tables without proper indexes the extraction performance can be dramatic.

    Suspect the following if loading to PSA is bad:

    - Datapackage reads data in small chunks (much smaller than 50000 records). Overhead causes more time than the actual data transport.

    - Network problems. Maybe the network is congested by other servers?

    If loading from PSA to datatarget is slow:

    - Check start routines for performance.

    - Are enough batch partitions available? Sometimes activation of ODS can be improved by more parallel processes.

    - Are you using a lot masterdata-lookups when filling the datatargets?

    When you report you findings in this forum we may be able to help you further.

    Add a comment
    10|10000 characters needed characters exceeded

Before answering

You should only submit an answer when you are proposing a solution to the poster's problem. If you want the poster to clarify the question or provide more information, please leave a comment instead, requesting additional details. When answering, please include specifics, such as step-by-step instructions, context for the solution, and links to useful resources. Also, please make sure that you answer complies with our Rules of Engagement.
You must be Logged in to submit an answer.

Up to 10 attachments (including images) can be used with a maximum of 1.0 MB each and 10.5 MB total.