When you need to retrieve archived data, follow this process to load it back into your BW system.
Key Concept
Archiving removes outdated data, but sometimes a need arises to analyze archived BW data and you must reload it back into the BW system. For example, your company archived and deleted order data from four years ago. The sales department now requests this data for new promotional analysis that it wants to perform. You cannot read this archived data directly into queries; instead, you must reload it back into BW and then analyze it in BW queries.
Once a BW system archives data, you may need to reload the archived data files back into the BW system because of legal or audit requirements. I’m going to walk you through the process you follow to reload BW archived data back into the BW source system where it originated. This data existed in a BW InfoProvider (InfoCube or ODS) and your company archived it to a file and deleted it from a BW source system.
The only two structures available for archiving are the InfoCube and ODS. Archiving is not available for master data or Persistent Staging Area (PSA) data. You can only reload BW data archived from the InfoCube or ODS back into BW.
I highly recommend that you not reload the archived data back into the original data target. For example, if you originally created an archived file from InfoCube A, you should not reload it back into InfoCube A. You should copy InfoCube A and create a new InfoCube B. You then can reload archived data into InfoCube B.
There are several reasons for this recommendation. The most important is that it allows the reloaded data to remain physically separate from any existing data. This allows for easier verification of the reloaded archive data because the data is isolated in its own InfoProvider. You also can combine reloaded data with current data via MultiProviders for better query performance because of the system’s parallel query processing functionality.
You should fully test the entire archival process in a development environment and then in the production environment. Therefore, you need to transport all objects that you created for this process using the BW transport functionality.
Note
Two distinct types of archived data exist: R/3 and BW archived data. R/3 archived data existed in R/3 and someone archived and deleted it from the R/3 system. BW archived data existed in BW and someone archived and deleted it from the BW system. Although in some cases you can directly load R/3 archived data into BW based on some standard business content extractors, this process is beyond the scope of this article. I will discuss the details about this process in
"Improve Your System Performance by Archiving Data in BW: Part 3." Load Archived Data
To see the archive information for an InfoProvider, go to the InfoProvider screen via transaction RSA1 and choose the InfoProvider tab. Right-click on an InfoProvider and choose Archive Administration… (Figure 1). The system brings you to the Archive Administration screen, where you can see a button labeled Management in Figure 2. This button displays all the archival runs for this InfoProvider and their status. If the Archive Administration option does not appear in Figure 1, this InfoProvider is not configured for archiving. To learn how to set up the InfoProvider for archiving, see my article, “Improve Your System Performance by Archiving Data in BW: Part 1.”
Figure 1
Archive Administration link from the InfoProvider
Figure 2
Archive Administration screen where you can access all archived data. The Management button allows you to see past archive runs and their statuses.
Archive administration tracks the previous archival runs to determine which archive files are relevant for your reloading needs. By expanding the information shown in the management screen, you can see the file name, date and time completed, status, number of records, and any notes that the system recorded during the archive run (Figure 3). Database table ADMI_FILES stores the information shown on this screen and keeps it in the system until someone deletes the InfoProvider.
Figure 3
Archive overview screen shows the archive run status and any notes that BW recorded when creating the archive
You are now ready to start loading the archived data back into an InfoProvider. First, create an export DataSource for the InfoProvider. This generates a DataSource and InfoSource that you can use to reload data. Right-click on the InfoProvider that you originally archived and click on Generate Export DataSource (Figure 4). The system creates an InfoSource and a DataSource.
Figure 4
Generate Export DataSource creates a DataSource and InfoSource to reload archived data
Next, replicate DataSources to make the system aware of the new DataSource you created. Go to the Administrator Workbench via transaction RSA1 and choose the Source Systems tab. Go to the BW source system. This source system is often called the MYSELF source system because it is the source system of the BW system itself. This system is usually first in the list of source systems.
Right-click on the MYSELF source system and click on Replicate DataSources. Now the export DataSource and related InfoSource are available. The system automatically assigns a prefix of 8 to any InfoSource used with an export DataSource. Thus, the name of the InfoSource is 8XXXXX, with XXXXX representing your InfoProvider name. You can view the InfoSource in the InfoSources tab of Administrator Workbench (transaction RSA1). Click on the binoculars icon to search and choose 8XXXXX.
To reload archived data into an Info-Provider, create a place to store the data. As stated previously, use a copied InfoProvider to keep this data separate. First, right-click on an InfoArea and then choose the Create InfoCube or Create ODS option (depending on the data’s destination). I am creating an InfoCube copy of InfoCube 0SD_C03 that I originally archived, so I enter the InfoCube name in the Copy from field (Figure 5).
Figure 5
Copy your existing InfoCube to load archived data
Now, establish update rules to fill this newly created InfoCube from the new InfoSource. Go to your InfoCube using the InfoProvider tab of Administrator Workbench, transaction RSA1. Select your InfoCube and choose Create Update Rules. Update the InfoCube via the 8XXXXX InfoSource. In this case, I’m using the InfoSource 8OSD_C03 (Figure 6). The copied InfoCube is ready to accept the reloaded archived data.
Figure 6
Create update rules from the 8XXXXX InfoSource to your copied InfoCube
Create the InfoPackage to use to load the archived data. Find your InfoSource in the InfoSources tab of Administrator Workbench (transaction RSA1). Right-click on the source system assignment under your 8XXXXX InfoSource and select the option to create an InfoPackage. BW creates an InfoPackage to use when reloading the archived data.
Once you complete archiving for an associated InfoProvider, a new button labeled Archiv.Sel appears on the bottom of the InfoPackage selection screen. The button allows you to choose the archived data files to load into the InfoProvider (Figure 7). Once you click on this button, three options appear:
- Database Only: BW loads only non-archived data into the InfoProvider. In other words, the InfoPackage acts like all other non-archive InfoPackages so that you can load data normally from the source system. This is the default setting that you should use when you do not plan to load archived data using this InfoPackage.
- Archive Only: BW loads only archived data via this InfoPackage. Click on the Archive File Select. button (in the pop-up box of Figure 7) to reload the archive file (Figure 8).
- Database and Archive: BW loads data from either of the two options above. If you choose an archive file, BW loads it; otherwise, it loads other source system data.
Figure 7
Options after you click on the Archiv.Sel button in the InfoPackage
Figure 8
Select the archived files for reload after clicking on the Archive File
I recommend that you choose the option Archive Only in the InfoPackage selection screen when loading archived data. This ensures that BW uses this InfoPackage for reloading archived data only. The selection screen in Figure 8 shows all archives that BW completed for this data target. This screen also shows any notes associated with this archive. You can go to Archive Administration for even more details on each archive (Figure 1).
Choose an archive to reload by checking the check box next to the archive. The system allows you to load multiple archives. After selecting the archives, using the tabs in the InfoPackage, complete the Processing screen and Data Targets screen of the InfoPackage to specify where and how to load the data.
When using a PSA to reload data, you can choose that option in the Processing screen. You select the destination of data in the Data Targets screen. Remember that the data target should not be the original source; it should be the newly created copied InfoProvider. You should see that listed as a possible data target because update rules exist from this InfoSource to this InfoProvider. BW only supports a full load of the archive data, so you cannot use a delta load. Execute the InfoPackage.
Note
You can preview the archived data before loading it by using Extractor Checker (transaction RSA3) on InfoSource 8XXXXX. Choose the archive file you wish to load and Extractor Checker shows a sample of the data that BW will load.
BW now reloads the data. If you are loading into an ODS, you can activate the ODS data immediately. InfoCube data is immediately available. You can now begin to verify and analyze the data using query processing, transaction LISTCUBE, or transaction SE16 to analyze the ODS tables. Depending on your query needs, you may need to create a MultiProvider to join your new archive only InfoProvider with any existing InfoProviders.
The system neither checks nor protects the InfoProviders from change after you create an archive file. For example, an InfoCube could contain the characteristics 0CUSTOMER, 0MATERIAL, and 0CALMONTH. You could create an archive file from this InfoCube, then you can add another characteristic like 0PLANT to the InfoCube. The archive file does not contain this characteristic, so BW does not populate any newly added characteristic once it reloads the data.
If an InfoProvider feeds other InfoProviders (data marts), the system does not automatically reload them after reloading archived data, so you must reload them manually. After loading the archived data, recreate or rebuild any aggregates associated with this data to aid query performance from the increased data volume.
Premium
