Write optimized parallel loads ?

Hello BI gurus,
Is parallel loads possible for write optimized ODS ?
if possible what are the limitations of it.
Please let me know your opinion.
Thanks in advance,
BWer

In write optimized DSOs the technical key is generated automatically when loading data, and the standard keys are not really required. If you still have those standard Keys , they are called "semantic keys". Those semantic keys can "bundled" into semantic groups in the DTP.
http://help.sap.com/saphelp_nw2004s/helpdata/en/b6/de1c42128a5733e10000000a155106/frameset.htm
Hope it helps,
Gili

Similar Messages

Write optimized dso - Stand DSO loading by Delta Req by req is not working

Hi All,
We have a data flow which has 1st layer having write optimized dso and then we are loading data to Standard DSO.
We are using Delta DTP to load the data with the option "Get All New Data Request By Request", but it is not working.
Suppose if i have 2 requests which are need to update to Standard Dso from Write Optimized Dso, it is updating in single request. We are on SAP BW 701 patch level is 005.
Please suggest me ....

Hi Prasanna,
For reasons of downward compatibility, the system behaves differently for DTPs that were created before "SAP NetWeaver 7.0 Support Package Stack 13"
DTPs for which the indicator is set only get the first new request, even if there is more than one new source request at the time of processing. This restricts the way in which these DTPs can be used in process chains, because requests accumulate in the source and the target may not contain the current data. The Retrieve Until No More New Data indicator is therefore displayed for these DTPs.
I Suggest you to set this indicator and activate the DTP.
If you set the Retrieve Until No More New Data indicator and then activate the DTP, once it completes processing, a DTP request checks whether there are any further requests in the source. If the source contains more requests, a new DTP request is automatically generated and processed. Once the DTP is activated, the indicator is no longer visible in DTP maintenance.
This applies when the DTP is started by a process chain and also when you start the DTP directly from DTP maintenance.
The indicator is set by default for new DTPs that get data request by request.
If you do not select the indicator, the label for the Get All New Data Request By Request indicator changes to Get One Request Only. Once the DTP is activated, only this indicator is displayed in DTP maintenance.
Regards,
Sudheer.

Error while loading data from write optimized ODS to cube

Hi All,
I am loading data from a write optimized ODS to cube
I have done Generate Export Datasource
schedulled the info packge with 1 selection for full load
then it gave me following error in Transfer IDOCs & TRFC
Info IDOC 1: IDOC with errors added
Info IDOC 2: IDOC with errors added
Info IDOC 3: IDOC with errors added
Info IDOC 4: IDOC with errors added
Data packege 1 arrived in BW Processing : selected number does not agree with transferred number
Processing below is green
shows update of 4 new records to Datapackage 1.
Please provide inputs for the resolution
Thanks & Regards,
Rashmi.

please let me know, What more details you need?
If I click F1 for error details i get following message
Messages from source system
see also Processing Steps Request
These messages are sent by IDoc from the source system. Both the extractor itself as well as the service API can send messages. When errors occur, several messages are usually sent together.
From the source system, there are several types of messages that can be differentiated by the so-called Info-IDoc-Status. The IDoc with status 2 plays a particular role here; it describes the number of records that have been extracted in a source system and sent to BI. The number of the records received in BI is checked against this information.
Thanks & Regards,
Rashmi.

Multiple data loads in PSA with write optimized DSO objects

Dear all,
Could someone tell me how to deal with this situation?
We are using write optimized DSO objects in our staging area. These DSO are filled with full loads from a BOB SAP environment.
The content of these DSO u2013objects are deleted before loading, but we would like to keep the data in the PSA for error tracking and solving. This also provides the opportunity to see what are the differences between two data loads.
For the normal operation the most recent package in the PSA should be loaded into these DSO-objects (as normal data staging in BW 3.5 and before) .
As far as we can see, it is not possible to load only the most recent data into the staging layer. This will cause duplicate record errors when there are more data loads in the PSA.
We all ready tried the functionality in the DTP with u201Call new records, but that only loads the oldest data package and is not processing the new PSA loads.
Does any of you have a solution for this?
Thanks in advance.
Harald

Hi Ajax,
I did think about this, but it is more a work around. Call me naive but it should be working as it did in BW3.5!
The proposed solution will ask a lot of maintenance afterwards. Beside that you also get a problem with changing PSA id's after the have been changed. If you use the posibility to delete the content of a PSA table via the process chain, it will fail when the datasourcese is changed due to a newly generated PSA table ID.
Regards,
Harald

Problem loading data into write optimized dso.....

Hi ,
I am having problem loading the data from PSA to write optimised DSO
I have changed the normal DSO into a Write Optimised DSO. I have 2 data sources to be loaded into the write optimized DSO. One for Demand and one for Inventory.
The loading of Demand data from PSA to DSO happens fine with out any error.
Now while loading the Inventory data from PSA to DSO , i get the below errors
"Data Structures were changed. Start Transaction before hand"
Execption CX_RS_FAILED Logged
I have tried reactivating the DSO, Transformation, DTP and tried to load the data again to the write optimised DSO, but no luck, I get the above error message always.
Can some one please suggest me what could be done to avoid the above error messages and load the data successfully.
Thanks in advance.

Hi,
Check the transformations is there any routines written.
Check your Data struture of cube and DS as well.
Is there any changes in structure.
Data will load upto PSA normally, if is there any changes in struture of DSO, then the error may occur.just check it.
Check the below blog:
/people/martin.mouilpadeti/blog/2007/08/24/sap-netweaver-70-bi-new-datastore-write-optimized-dso
Let us know status.........
Reg
Pra

Duplication Error while loading data in write optimized DSO

Hi Experts,
I have an issue.In BI7 I'm trying to load the data in a WRITE OPTIMIZED ODS from the controlling data source 0CO_OM_CCA_10. I'm getting the data properly in PSA, but while loading it into my WODSO i'm getting the duplication error although my keys fields and data fields are properly placed in my data target(WODSO).
pls let me know what is the solution to load it successfully.
Thanks in Advance.
Amit

Hi,
thanks for your reply
I'm getting this error message:
Diagnosis
    During loading, there was a key violation. You tried to save more than
    one data record with the same semantic key.
    The problematic (newly loaded) data record has the following properties:
    o   DataStore object: GWFDSR02
    o   Request: DTPR_4BA3GF8JMQFVQ8YUENNZX3VG5
    o   Data package: 000006
    o   Data record number: 101
Although i have selected the key fields which identifies unique record,then also i'm getting the duplication error.
Even i have reffered to the BI content for this data source and found that it has the same key fields as of mine.
Debjani: i need unique records without duplication and i'm doing a full load in DTP.
What is to be done pls help
Thanks in advance.
Edited by: Amit Kotwani on Sep 26, 2008 10:56 AM

Duplicate Error while loading data to Write Optimized DSO

Hi,
When i do a dataload for Write Optimized DSO, I am getting an Error as "Duplicate Data Record Detected". I have Sales Document Number, Fiscal Year Variant & Billing Item as Semantic Key in the DSO.
For this DSO, I am getting data from a Test ECC system, in which most of the Sales Document Number column is Blank for this Datasource.
When i go into the Error Stack of the DSO, all the rows whichever has Sales Document number as Blank are displayed. For all this rows, the Item Number is 10.
Am i getting this Duplicate error as the Sales Document Number is Blank & Item number is 10 for all of them? I read in Threads that Write Optimized DSO doesnt care about the Key Values, it loads the data even if the Key values are same.
Any help is highly appreciated.
Regards,
Murali

Hi Murali,
Is the Item Number a key field ?
When all the key fields are same then data gets aggreagted depending on the setting done in the transformation for the key figures. These 2 options for key figures are :
1. Add up the key figures
2. Replace the key figure
Since the Sales Document No is blank and Item Number is same then their is a possibility that the key figures for these records might get added up or replaced and thus due to this property of SDSO it might not be throwing error.
Check the KF value in the SDSO for that Sales Doc No and Item No. and try to find out what is the value for KF. It may have addition of all the common data fields or the KF value of last common record.
Regards
Raj Rai

Error during loading and deletion of write-optimized DSO

Hey guys,
I am using a write optimized DSO ZMYDSO to store data from several sources (two datasources and one DSO).
I have disabled the check of uniqueness in the DSO, but I defined a semantic key for the fields ZCLIENT, ZGUID, ZSOURCE, ZPOSID which are used in a non-unique index.
In the given case, I want to delete existing rows in the DSO. I execute these steps in the endroutine. Here the abstract coding:
LOOP AT RESULT_PACKAGE ASSINING <RESULT_FIELDS>.
u201Csome other logic [u2026]
DELETE /BIC/AZMYDSO00
WHERE /BIC/ZCLIENT = RESULT_FIELDS-/BIC/ZCLIENT
     AND /BIC/ZGUID = RESULT_FIELDS-/BIC/ZGUID
AND /BIC/ZSOURCE = RESULT_FIELDS-/BIC/ZSOURCE
AND /BIC/ZPOSID = RESULT_FIELDS-/BIC/ ZPOSID.
ENDLOOP.
COMMIT WORK AND WAIT.
During the Loading (after the transformation step in the updating step), I get the messages (not every time):
1.     Error while writing the data. (RSAODS131)
2.     Could not Save DataPackage xy in DataStore ZMYDSO (RSODSO_UPDATE027).
Diagnosis: DataPackage XY could not be saved. Reasons therefore could be violation of key uniqueness (duplicate data) or general database error.
3.     Error in the substep of updating DataStore.
I have checked the system log (SM21) and the system dumps (ST22) but I could not find an exact error description.
I guess, I am creating some inconsistencies or locks (I also checked the SM12) so that the load process interrupts. But I also tried a serial updating within the DTP (I reduced the number of batch processes to 1). No success.
Perhaps the loading of one specific package could take a longer time so that the following package would overtake the predecessor. Could that be a problem? Do you generally advise against the deletion of rows within the endroutine?
Regards,
Philipp

Hi,
is ZMYDSO the name of the DSO?
And is this the end routine of the transformation while loading the same DSO?
if so we never do such a thing.
you are comparing the DSO with the data that is flowing in and then deleting the data from the DSO...
Which doesnt actually make any sense... because when loading the data to a DSO (or a cube or any table) the DSO (or cube) will be locked exclusively for any modifications of data. You can only read data from it.
If your requirement is that existing duplicate records need not arrive in the DSO then you can delete the data from the SOURCE_PACKAGE in the start routine like below
SELECT FIELDS FROM /BIC/AZMYDSO00 INTO INTERNAL_TABLE WHERE <CONDITION>.
LOOP AT INTERNAL_TABLE.
   DELETE SOURCE_PACKAGE
WHERE SOURCE_PACKAGE-/BIC/ZCLIENT = INTERNAL_TABLE-/BIC/ZCLIENT
     AND SOURCE_PACKAGE-/BIC/ZGUID = INTERNAL_TABLE-/BIC/ZGUID
AND SOURCE_PACKAGE-/BIC/ZSOURCE = INTERNAL_TABLE-/BIC/ZSOURCE
AND SOURCE_PACKAGE-/BIC/ZPOSID = INTERNAL_TABLE-/BIC/ ZPOSID.
ENDLOOP.
or if your requirement is that you need to delete the old data from the DSO for the same key which is arriving newly in order to load the new data into the DSO in that case, you could do something like this in the start routine
SELECT FIELDS FROM /BIC/AZMYDSO00 INTO INTERNAL_TABLE FOR ALL ENTRIES IN SOURCE_PACKAGE
WHERE /BIC/ZCLIENT = SOURCE_PACKAGE-/BIC/ZCLIENT
     AND /BIC/ZGUID = SOURCE_PACKAGE-/BIC/ZGUID
AND /BIC/ZSOURCE = SOURCE_PACKAGE-/BIC/ZSOURCE
AND /BIC/ZPOSID = SOURCE_PACKAGE-/BIC/ ZPOSID.
* now update the new values you want to write in the loop
LOOP AT INTERNAL_TABLE INTO WORK_AREA.
"CODE FOR MANIPULATION of WORK_AREA
*write a modify statement to update the RESULT_PACKAGE.
MODIFY RESULT_PACKAGE FROM WORK_AREA TRANSPORTING FIELDS.
ENDLOOP.
hope it helps,
Regards,
Joe

Delta loading procedure from Write Optimized DSO to Infocube

Hi All,
We are using Write optimized DSO in our project to which I am loading data using Standard DSO 0FI_GL_12
From Write optimized DSO, we are loading delta records into Infocube, please provide your inputs to my following questionnaire:
1) I am quite interested to know how the delta records gets loaded into infocube whenever we are using 'Write optimized DSO' as we don't have any image concept in Optimized DSO
Ex: If am using Standard DSO, we have Change log table and image concept will allow to get updated value to Cube
let us assume
Active Table
111            50
111            70 (overwrite)
Change Log Table
111            -50        (X -- Before Image)
111             70    ( ' ' -- After Image) symbol for after image is 'Space'
So if we load this record to the target as a delta the above two records from change log table will get loaded to the CUBE and Cube will have 70 as the updated record
If am using 'Write Optimized',
Active Table
111            50
111            70 (overwrite)
When this record loaded to the cube, as Info Cube is always having 'Additive' feature so the total value will 50+70 =120 which is wrong?
Correct me what feature will work here to get updated value as '70' into Cube from Write Optimized DSO'
2)As the data source is delta capable and having 'ADDITIVE' delta process, only the delta records based on REQUEST ID are loaded into Info Cube with the updated key figure value?
Thanks for your inputs and much appreciated.
Regards,
Madhu

Hi Madhu,
In best practice, we use WODSO in Initial layer and then Standard DSO. Just for mass data load/staging purpose.
In best practice : Data source ----> WODSO ---> std. DSO
In your case : Data source ----> Std.DSO -----> WODSO.
In both cases if data load design is not in accurate way, then your cube will have incorrect entries.
For ex: today 9 am : 111, 50 (in active table)
Data load to cube, same day 11 am : then cube will have 111    50.
Same day, value got changed in std. DSO 1 pm :   111   70(over write function @ active table).
Same day/next day if you load data to cube, it will have 2 records one is with value 50 and other would be 70. So to avoid such scenarios we should plan load in accurate way. Else make change your DTP settings as ‘source table : change log table.
Coming to your case:
Once after the load to Std. DSO, load data to WODSO by changing the DTP settings ‘Delta Init.Extraction from’ : Change log.
Now data available @WODSO from change log table, then you load to cube In delta mode.

Load Data with 7.0 DataSource from Falt file to Write Optimized DSO

Hi all,
we have a problem loading data from flat file using the 7.0 datasource.
We have to load a flat file (monthly) into a WO DSO. The infopackage load file in full mode into the Datasource (PSA) and the DTP Load in delta mode data from datasource into the WO DSO.
When i load the second file in the Datasource, the DTP load all data present in the Datasource and not only the new one as aspected using Delta mode.
Has anyone any tips to help me?
Thank you for help.
Regards
Emiliano

Hi,
Iam facing the similar problem.
Iam using Write Optimized DSO and i have got only 1 req in PSA (have deleted all previous req from PSA and DSO).
When iam doing a delta load from PSA to DSO, i expect to see only that 1 req to get loaded into DSO.
But, its picking up the data from 3 other reqests and doubling the records...
Can you please help me, how did you managed to get out of that isue?
Cheers,
Nisha

Load performance Write-Optimized DSO

Dear all,
I'm looking for some practical tips concerning improving load performance from the PSA to a write-optimized DSO (41M records) via a DTP.
All parameters that could be tweaked have been checked (e.g. packet size, batch jobs, uniqueness of data flag, etc.) and optimized.
However this load stays extremely slow (init load from source to BW is faster).
The BW system runs on SP18.
Please share all your tips our recommendations for us to solve this major issue.
Your help is much appreciated!
Thanks
JvB

Hi JvB,
Here are some options I can think of and let you know if I remember something:
1) Increase number of prallel processes in DTP i.e. DTP->GOTO->Settings for bacth monitor->Number of prallel process to 4 or 5.
2) Note 409641 - Examples of packet size dependency on ROIDOCPRMS
The general formula for data transfer is:
packet size = MAXSIZE * 1000 \ transfer structure size
but not more than MAXLINES.
eg. if MAXLINES < than the result of the formula, MAXLINES size is transferred into BW.
3) you should check if there are any locks or deadlocks in ST04 and also System Analysis SM21....it will show u all the details,,, and Short dump in ST22.....there may be some ..
4) Check and Analyze DSO in RSRV. If there are any issues repair it.
5) If you are loading huge volumes of data split the records by Filters in DTP ex. Plant, calendar year or calendar month selections.
And also Try
Try partioning,craeting indexes/secondary indexes or archiving old data.
Check these links for furtehr reference:
/people/martin.mouilpadeti/blog/2007/08/24/sap-netweaver-70-bi-new-datastore-write-optimized-dso
/message/2987899#2987899 [original link is broken]
Good Luck.
Regards
Satish Arra
Edited by: Satish Arra on Feb 7, 2009 9:28 PM

Disk throughput drops when inserting data packages in write-optimized DSO

Hi all,
we are currently testing our new freshly installed SAN.
To see the performance gain in BI, I'm currently doing some test loads.
And during the monitoring of those loads, I noticed something I'd like someone to explain :-):
I execute a DTP from PSA to a write-optimized DSO.
The n° of parallel processes = 9
Update method = serial extraction, immediate parallel processing
N° of records transferred: +23.000.000
Ok, in the first phase (read the PSA) only one process is used (serial extraction). When I look in OS07, I notice we have very good throughput: +66.000 TransfKB/s. Very nice!
But as soon as BI starts inserting the data packages, and parallel processing kicks in, the throughput drops to 4K or something, and sometimes we get 20K at max. That's not too good.
We have a massive SAN , but the BI system does not seem to use it?
I was wondering why this is the case. I already toyed around with the package size, but it's always the same.
Also I noticed that the allocated processes don't seem to be active. I allocated 9 BTC processes to this load.
They are all used, but we only see 3 inserts at the same time, max. Also in the DTP-monitor, only 3 packages are processed at the same time. As it's a write-optimized DSO, RSODSO_SETTINGS does not apply I presume.
Any ideas?
tnx!

Hi,
can you pls try to give some filetr in DTP and try to pull the data.
I am not sure why first data package is taking long time and otehr data package is taking less time..
Do you have any start routine..If datapak = 1.. the do this logic..
Pls check..
regards
Gopal

Changes to write optimized DSO containing huge amount of data

Hi Experts,
We have appended two new fields in DSO containg huge amount of data. (new IO are amount and currency)
We are able to make the changes in Development (with DSO containing data). But when we tried to
tranport the changes to our QA system, the transport hangs. The transport triggers a job which
filled-up the logs so we need to kill the job which aborts the transport.
Does anyone of you had the same experience. Do we need to empty the DSO so we can transport
successfully? We really don't want to empty the DSO's as it will take time to load?
Any help?
Thank you very muhc for your help.
Best regards,
Rose

emptying the dso should not be necessary, not for a normal dso and not for a write optimized DSO.
What are the things in the logs; sort of conversions for all the records?
Marco

4 LSA architecture - EDW layer - write optimized DSO settings

Dear Colleagues,
I have a question regarding the 4 LSA architecture, available on the following article.
http://www.sdn.sap.com/irj/scn/index?rid=/library/uuid/306f254c-1e3f-2d10-9da0-bcff4e35e0ef
When we activate a SAP business content dataflow.
If we want to add in the EDW layer write optimized DSO to have a faster upload from source system to SAP BI.
Should we always check the" do not check uniqueness of data" setting in the write optimized DSO to avoid data upload error ?
Based on your experience what would be your recommendation ?
Cheers,

I would suggest to check "ON"
-If the check is ON It will allow to load several records with same semantic key and in next layer
-If the check is OFF It will not allow to load same record twice and throws error
Did you seen this ?
/people/martin.mouilpadeti/blog/2007/08/24/sap-netweaver-70-bi-new-datastore-write-optimized-dso
Edited by: Srinivas on Aug 24, 2010 2:16 PM

Duplicate Semantic Key in Write Optimized DSO

Gurus
Duplicate semantic keys have a unique index KEY in the key fields of the DSO when WRITE OPTIMIZED DSO is used. (Of course this is assuming the Do not check uniqueness of data indicator is not checked..)
See help https://help.sap.com/saphelp_crm60/helpdata/en/a6/1205406640c442e10000000a1550b0/frameset.htm
This means that the DSO can contain duplicate records.
My question is: What happens to these duplicates when a request level delta update is done to a Standard DSO or Infocube?
Do duplicates end up in the error stack? Are they simply aggregated in further loads? - because this would be a problem for reporting (double-counting).
thanks
tony

Hi Tony,
It will aggregate the data in some undesired way.
Read on...
https://help.sap.com/saphelp_crm60/helpdata/en/b6/de1c42128a5733e10000000a155106/frameset.htm
If you want to use write-optimized DataStore objects in BEx queries, we recommend that they have a semantic key and that you run a check to ensure that the data is unique. In this case, the write-optimized DataStore object behaves like a standard DataStore object. If the DataStore object does not have these properties, unexpected results may be produced when the data is aggregated in the query.
Hope it helps...
Regards,
Ashish

Write optimized parallel loads ?

Similar Messages

Maybe you are looking for