Skip to Content

Have you ever thought: How many steps do I need to do, until I can load a flat file into BW? I need so many objects, foremost I need lots of InfoObjects before I can even start creating my infoprovider. Then I need DataSource, Transfomation (where I need to draw many arrows), DTP and InfoPackage. I just want to load a file! Why is there no help by the system?

Now there is – BW 7.30 brings the DataFlow Generation Wizard!

You start by going to BW DataWarehouse workbench (as always), then selecting the context menu entry „Generate Data Flow…” on either the File-Source system (if you just have a file and want to generate everything needed to load it) or on an already existing DataSource (if you have that part already done – this works also for non-file-source systems!) or on an infoprovider (if you have your data target already modeled and just want to push some data into it).

Context Menu to start data flow wizard

Then the wizard will pop up:

Step 1 - Source options

Here, we have started from the source system. If you start from the InfoProvider, the corresponding step will not be shown in the progress area on the left, since you have selected that already. Same for the DataSource.

I guess you noticed already: ASCII is missing in the file type drop down (how sad! – however please read the wizard text in the screenshot above: It’s just the wizard where it is not supported because the screen would become too complex). And look closer: There is „native XLS-File“. Yes, indeed. No longer „save as CSV“ necessary in Excel. You can just specify your Excel-File in the wizard (and in DataSource maintainance as well). There is just one flaw for those who want to go right to batch upload: The Excel installation on your PC or laptop is used to interpret the file contents, so it is not possible to load Excel files from the SAP application server. For this, you still need to save as CSV first, but the CSV structure is identical to the XLS structure, so you do not need to change the DataSource.

Ok, lets fill out the rest of the fields, file name of course, data source, source system, blabla – (oops, all this is prefilled after selecting the file!) – plus the ominous Data Type (yes, we still can’t live without that)

Step 1 - Pre-Filled Input Fields

 

and „Continue“:

Step 2 - CSV Options

 

Step 2 - Excel Options

 

One remark on the header lines: If you enter more than one (and it is recommended to have at least one line containing the column headers), we expect the column headers to be the last of the header lines, i.e. directly before the data. Now let‘s go on:

Step 3 - Data Target

 

The following InfoProvider Types and Subtypes are available:

  • InfoCube – Standard and Semantically Partitioned
  • DataStore-Object – Standard, Write Optimized and Semantically Partitioned
  • InfoObject – Attributes and Texts
  • Virtual Provider – Based on DTP
  • Hybrid Provider – Based on DataStore
  • InfoSource

This is quite a choice. For those of you which got lost in that list, have a look at the decision tree which is available via the „i“ button on the screen. As a hint: A standard DataStore-Object is good for most 😉

Step 4 - Field Mapping

This is the core of the wizard. At this point, the file has already been read and parsed, and the corresponding data types and field names have been derived from the data of the file and the header line (if the file has one). In case you want to check whether the system did a good job, just double click the field name in the first column.

This screen does also define the transformation (of course only 1:1 mapping, but this will do for most cases – else you can just modify the generated transformation in the transformation UI later) as well as the target infoprovider (if not already existing) plus the necessary InfoObjects. You can choose from existing InfoObjects (and the „Suggestion“ will give you a ranked list of InfoObjects which map your fields better or worse) or you can let the Wizard create „New InfoObjects“ after completion. The suggestion uses a variety of search strategies, from data type match via text match to already used matches in 3.x or 7.x transformations.

And that was already the last step:

Step 5 - End

After „Finish“, the listed objects are generated. Note, that no InfoPackage will be generated, because the system will generate the DTP to directly access the file rather than the PSA.

 

Legal Disclaimer

Don’t miss any of the other Information on BW 7.30 which you can find SAP Business Warehouse  7.3

To report this post you need to login first.

15 Comments

You must be Logged on to comment or reply to a post.

  1. Peter Baumann
    Hello!

    Thank you for posting this. First, I like this because it makes the BW more handy and other tools in the BI area has similar functions since years.

    I don’t want to be to negative but I think this can only be used for testing or for simple cases. From my experience SAP BW is not very often used for this simple cases… how does this functionality fits in concepts a corporate BI architecture? I don’t know but time will show.

    sometimes I think with new releases there come more and more single functions. Maybe this is a typical german problem. We build cars which can do everything if you find out how. I had driven some french cars which have much less functions. But the functions the have are very intuitive to use. That is what I really like. Not just have a amount of functions and need a week of training to learn how to use them. Sit in and use them immedately, without thinking. But think of the higher concept. In the last consequenze the car should drive. And that efficently for a longe time.

    At least I’m curious what else comes. Let’s go…

    Best regards,
    Peter

    (0) 
    1. Thomas Rinneberg Post author
      Hi!

      Thanks for your reply! You are right in saying, that this wizard covers only a small portion of what needs to be modeled for an enterprise data warehouse. For that reason, it is embedded also into the 7.30 dataflow template modeler. Curious? Hopefully soon a colleague will write a weblog about this new tool as well.

      Regarding features, I would say that 7.30 is the TCD/TCO release, since it first time in history provides more features that make your live easier than features which add even more capabilities. Examples are the semantically partitioned objects, the Business Objects Data Services Modeling Integration, or the data flow copy wizard.

      Best Regards, Thomas

      P.S.: I drive a French car as well 😉

      (0) 
  2. Kiran Subramaniam
    Now any new BW learner does not need to understand the intricacies of transformations, DTP’s, Infopackages and PSA . Load data in 5 min and create a report – Zero to WOW in 5 minutes time slot 🙂

    Thank you,
    Kiran

    (0) 
  3. Bhanu Gupta
    Sometimes it does feel like too many steps that you need to go through to load a simple file even when you don’t need any logic to modify the data. This looks interesting.
    (0) 
  4. Witalij Rudnicki
    I was really excited about this new BW wizard and how this can facilitate building LSA-based BW. But after playing with it my wish list is pretty long (based on SP 0):
    – I still think wizard can be more complex, like e.g. one from SPO
    – Wizard should recognize units/currencies missing in DataSource and still add them to the target structure with some default values
    – Have date fields recognized as Time Characteristics, not as regular Characteristics
    – Have ability to add Time Characteristics to the Key Fields of target DSO
    Regards,
    -Vitaliy
    (0) 
    1. Thomas Rinneberg Post author
      Hi!

      Thank you for your feedback!

      – The main focus of this wizard is simplicity and time to result. The SPO wizard in turn requires some time to get used to all it’s possibilities.
      – If a number stands without units / currencies in the file, the system cannot know that it is an amount or quantity. But via double click on the field in the transform step, you can change the type to CURR or QUAN. Then you will also be prompted to enter the related CUKY or UNIT field or to enter a fixed unit by enclosing it in single quotation marks (e.g. ‘USD’).
      – If the data type of the field is DATS, 0CALDAY is the InfoObject of choice to be proposed. It might however be that the field is not recogniced as date in the first place due to some unusual date format. You might change this as well in the transform step but probably need a special conversion exit then.
      – Your last point will come with patch 2, cf. note 1498389 as soon as released for all changes to the wizard in patch 2.

      Best Regards, Thomas

      (0) 
  5. Shani Bashkin
    This sound like a great helpful tool, especially to  test case scenarios and other sides scenarios to build in a second.
    I think it will be a great help to juniors or basis consultants who need to create a flow example in order to check or understand BW some issues.

    I didn’t completely understand if the wizard creates all the relevant objects as regular objects, so we will be able to modify it afterwards as we wish.
    Can you tell if this is the case?

    Thanks a lot,
    Shani

    (0) 
  6. Shani Bashkin
    This sound like a great helpful tool, especially to  test case scenarios and other sides scenarios to build in a second.
    I think it will be a great help to juniors or basis consultants who need to create a flow example in order to check or understand BW some issues.

    I didn’t completely understand if the wizard creates all the relevant objects as regular objects, so we will be able to modify it afterwards as we wish.
    Can you tell if this is the case?

    Thanks a lot,
    Shani

    (0) 

Leave a Reply