on 11-16-2005 2:07 PM
Hi experts,
we're currently trying to extract sales statistics from IS-U into BW. Using the extractorchecker (RSA3) the data (especially the number of records) seems to be correct. When initializing the extraction from within BW, we only get a fraction of the data??!! As far as I know, this should be the same. Are we missing something, customizing, parameters, etc....?
Thanks in advance.
Kind regards,
Micky Oestreich
Hi Micky,
Probably you r using initialization with diffrent selection criteria. Check in BW side, whether you are extracting all data to BW .i.e. different infopackages with different selection criteria. If you request for all the Data to BW, it should come. Where you r checking the is it PSA or Data Targets.
Elaborate wht u have done for extracting data to BW. I will have a look into it. Let me know if i'm wrong.
all the best.
Nagesh.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi Scott & Nagesh,
Thanks for your reply. With respect to your questions I have the following remarks:
We compared the data in BW using the content of the PSA table (= an exact image of the extracted data)
In RSA3 we used the F(full update setting)
In BW we executed FULL upload.
Both selections were the same (only one option availble: posting date). Period = 20050101 - 20051231 in both cases.
During our analysis, we found out that the SBIW setting "Maintain control parameters for data transfer" in IS-U has an substantial effect on the amount of extracted records!!
The parameter "Max lines per package" determines somehow HOW MANY records are extracted to BW.
When we set this parameter to 1, all records are extracted but this leads to 1-60 records per package -> we get a humungous amount of packages in the BW monitor -> performance is really, really poor....
How is this possible???
Thanks & regards,
Micky
Micky,
The Data Transfer control settings are specifically meant to have a profound effect on the way data is extracted based upon the source system, but should not effect the total <b>NUMBER</b> of records returned. The notes in SBIW outline exactly how each field within a record of that table effects the data transfer process. See Below:
1. Source System
Enter the logical system of your source client and assign the control parameters you selected to it.
You can find further information on the source client in the source system by choosing the path Tools -> Administration -> Management -> Client Maintenance.
2. Maximum Size of the Data Packet
When you transfer data into BW, the individual data records are sent in packets of variable size. You can use these parameters to control how large a typical data packet like this is.
If no entry was maintained then the data is transferred with a default setting of 10,000 kBytes per data packet. The memory requirement not only depends on the settings of the data packet, but also on the size of the transfer structure and the memory requirement of the relevant extractor.
3. Frequency
The specified frequency determines the number of IDocs that an Info IDoc is to be sent to, or how many data IDocs an Info Idoc describes.
Frequency 1 is set by default.. This means that an Info Idoc follows every data Idoc. In general, you should select a frequency between 5 and 10 but no higher than 20.
The bigger the data IDoc packet, the lower the frequency setting should be. In this way, when you upload you can obtain information on the respective data loading in relatively short spaces of time .
With the help of every Info IDoc, you can check the BW monitor to see if there are any errors in the loading process. If there are none, then the traffic light in the monitor will be green. The Info IDocs contain information such as whether the respective data IDocs were uploaded correctly.
4. Maximum number of parallel processes for the data transfer
An entry in this field is only relevant from release 3.1I onwards.
Enter a number larger than 0. The maximum number of parallel processes is set by default at 2. The ideal parameter selection depends on the configuration of the application server, which you use for transferring data.
5. Batch job target system
Enter the name of the application server on which the extraction job is to be processed.
To determine the name of the application server, choose Tools -> Administration -> Monitor -> System monitoring -> Server. The name of the application server is displayed in the column Computer
If you are experiencing inconsistencies still please give more detail and I will try to assist.
Cheers,
Scott
Hi Micky,
The only difference between the two scenarios that I see you are describing is the Update Mode. When running transaction RSA3 by default the update mode is set to F, or Full Mode. Update mode C, or Delta initialization which is what you are trying to use is obviously bringing you back the incorrect results. Are you testing in RSA3 with the appropriate update mode?
If that is the case then where exactly are you checking and finding your inconsistency? in the PSA or a lower level ODS Object.
Remember if there is any type of summarization within the data layers of your DataWarehouse you will receive less records. Make sure that your extraction structure and the transfer structure are identical for reconciliation purposes, once you have done this re-run the load and let me know if you are still receiving the same issue.
I hope this helps a bit.
Cheers,
Scott
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
User | Count |
---|---|
61 | |
9 | |
7 | |
6 | |
6 | |
6 | |
5 | |
5 | |
5 | |
5 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.