Operational analytics with SAP HANA Live - including batch attributes & variant configuration

[stefan_weisenberger]

The possibility of real-time, flexible and ad-hoc operational analytics is for many customers a prime motivation for moving to SAP Business Suite on HANA. Combine this with a beautiful self-service tool like Lumira and enable your business users to explore & anaylze themselves. Just this is already a big value prop in itself.
What are the ingredients?

• First, you need either to have your ERP or Business Suite on SAP HANA, or run HANA Analytics in a side car mode.


• With this comes SAP HANA live - if you want to explore, loot at this SCN post for starters. There is an abundance of documents in SCN on the topic. For example: How to Enhance SAP HANA Live - Webinar Presentation http://scn.sap.com/docs/DOC-46825


• And yes - there is of course a rapid deployment solution (RDS) available - with recent updates like "how to realize profitability analysis (CO-PA) based on SAP HANA Live", or a pre-assembled appliance for SAP HANA Live to be deployed in SAP HANA Enterprise Cloud (HEC).


• Finally, you connect it to SAP Lumira to make your data look beautiful and easily accessible- a few good examples for starters:  http://scn.sap.com/docs/DOC-55966 - or check below Lumira screenshot of an analysis of configured sales orders.

 



There are many good descriptions of how to to set this-up. I suggest the RDS - this should get you started, well, rapidly.



 

What about attributes, characteristics, configurations and classifications?

As always mill products we talk about rolls and sheets, every sales order and stock item is somewhat different - modeled through attributes or characteristics. Are those supported in SAP HANA live as well?

 

The short answer: yes, but you will need to "enable" them.

 

For many of the "normal" data base tables SAP delivers predefined HANA Live content - virtual model definition, like views. For classification and configuration the process is slightly different. Every customer can define their own attributes to describe a demand (e.g. sales order) or supply (e.g. a batch in stock) - as part of variant configuration and classification.

As we from SAP do not know your characteristic definitions in advance, we can't deliver ready to run HANA live views for those. Instead you will need to run a report to generate the HANA live views for classification and configuration in your system.
The result is similar: The view with your attributes is then available in your HANA content. From there you can start to link e.g. a configuration with a sales order query view, or with a inventory view for batches. Looking for a batch stock overview to search and display by attributes. Here you go.

 

Technical details & how-to

• Classification - see SAP OSS notes 1820261, 1892589 and 1826411


• SAP online help on SAP HANA Live model for characteristics & characteristic descriptions (which also explains how the technical table fieldnames are "translated" into a simplified & easier understandable virtual data model


• Variant Configuration - see SAP OSS notes 1987669 and 1983086


• See the RDS - the activation of the classification views is part of the RDS. Look for the scope option 5 or SAP HANA Live for SAP Product Lifecycle Management.

 

Let us know about the analytical models you have built, and share with the community your best findings!

[former_member204296]

stefan.weisenberger,

One question, any idea about how to run analytics on saved configurations in SAP Lumira? We have sales orders saved as xml files and stored in a compressed foramt, wanted to run similar analytics.

Thanks

Jagannatha

[stefan_weisenberger]

Hi Jagannatha,

just to clarify - are we talking about the same thing? (I am not the big Lumira expert.)

My focus is here on SAP ERP's variant configuration or batch classification, and how to represent the ERP data model in a simplified way in SAP HANA Live.

If I understand you correctly, you are working with a different data source, right?

I suggest to ask this question in the SAP Lumira space. You are more likely to get an expert answer on this question there.

Shot from the hip, you may want to take a look at

The specified item was not found.

Regards,
Stefan

[Vaishalibabar]

Hi Stefan,

Nice blog.

I have question over here , if we have 50 classes does that mean 50 Classification views we have to generate with the Program suggested over in the SAP Note. Do we have single view of all the Material its associated class and characteristic which belongs to it which we can plugin to any query for further use.

Regards,

Vaishali.

[stefan_weisenberger]

Hi Vaishali,

thank you for your smart question. I have not responded yet as I am still in discussion with our development team about the best approach.

The implementation of HANA live for classification and configuration is slightly different. Are you asking about classification, or variant config?

Let me get back to you later this week.

Thanks and regards,

Stefan

[Vaishalibabar]

Hi Stefan,

I am talking about HANA live for Classification only.

I have scenario where I want to pull all the 023 Classes and its associated characteristic and their values.

We have near about 40 Batch Class available in ECC.

So I was thinking do we need to created that many views so that we can plugin to any model.

Having 40 views and then plugin to any model would really make the model cumbersome.

So I was trying to seek an approach as how we can proceed.

Thanks for your prompt response and I would wait to hear back from you.

Regards,

Vaishali.

[stefan_weisenberger]

Dear Vaishali,

I have discussed this with our development team.

As far as we understood you’re trying to have all of the information (characteristics and characteristic values) of the 023 classes in your system available through one big ‘interface’ that you can plug to your classified objects. That’s a nice idea from a reuse perspective because like this you would only have one big data model.

However, we followed a different approach when generating the calculation views. We designed the generation report in a way that it helps to answer questions for specific Use Cases (by generating them specifically for one class). Like this, we keep the result set smaller (fewer columns) and therewith more tailored to the specific Use Case / Question.

This is also due to the fact that a column is created for each characteristic of a class. For each column that is created, a JOIN is performed by the calculation view. If you now consider all characteristics of all classes this would mean really a lot of columns (and JOINS) out of which just a minimal subset would be interesting for the User.

So, if you have specific questions to answer try to use specific calculation views that just contain the necessary classes.

Besides of that, you’re right. If you need information regarding characteristics and characteristic values for all classes at the same time this is a bit of work to do.

We have also discussed the option to alternatively manually create one "selection class" that would contain all characteristics relevant for reporting. But this would not work out-of-the-box.

The generated calculation view always contains the CLINT, the internal class number.

After generation, you would need to manually add the other CLINTs to the generated calculation view.

Let me know if this helps, and if you have additional questions.

Thanks and best regards,

Stefan