on 05-14-2014 7:06 PM
Hello Experts,
We are in the process of HANA Standalone implementation and design studio as reporting tool. When I am modeling, I did not figure out answers to some of the below questions .Below are the questions. Experts, please help.
Best way of modeling: The SAP HANA LIVE is completely built on calculation view; there are no Attribute and Analytical views. I have got different answer why there is only Calculation view and there are no Alaytic view and Attribute views. We are in SP7 latest version. This is a brand new HANA in top of non-SAP (DB2 source). What is the best way to model this scenario, meaning, can we model everything in the Calculation view’s like SAP HANA live or do you suggest using the standard attribute, analytical and calculation views to do the data model. Is SAP moving away from AV & AT to only calculation Views to simply the modeling approach?
Reporting: We are using the design studio as front end tool. Just for example, if we assume that we are
Using the BW, we bring all the data in to BW from different sources, build the cubes and use the bex query. Here in bex query we will be using the restricted key figures, calculated key figures calculations etc. From the reporting wise, we have the same requirements, calculations, RKF, CKF,Sum, Avg etc. if we are Using the design studio on top of standalone HANA, where do I need to implement all these calculations? Is it in different views? (From reporting perspective, if it’s BW system, I would have done all the calculations in BEx.)
Universe: If we are doing all the calculations in SAP HANA like RKF. CKF and other calculations , what is the point in having additional layer of universe , because the reporting compnets cam access the queries directly on views .In one of our POC , we found that the using universe affect performance.
Real time reporting: Our overall objective is to give a real time or close to real time reporting requirements, how data services can help, meaning I can schedule the data loads every 3 or 5 min to pull the data from source. If I am using the Data services, how soon I can get the data in HANA, I know it depends on the no of records and the transformations in between the systems & network speed. Assuming that I will schele the job every 2 min and it will take another 5 min to process the Data services job , is it fair to say the my information will be available on the BOBJ tools with in 10 min from the creation of the records.
Are there any new ETL capabilities included in SP7, I see some additional features included in SP7. Is some of the concepts discussed are still valid, because in SP7 we have star join concept.
Thanks
Magge
magge kris wrote:
Hello Experts,
We are in the process of HANA Standalone implementation and design studio as reporting tool. When I am modeling, I did not figure out answers to some of the below questions .Below are the questions. Experts, please help.
Best way of modeling: The SAP HANA LIVE is completely built on calculation view; there are no Attribute and Analytical views. I have got different answer why there is only Calculation view and there are no Alaytic view and Attribute views. We are in SP7 latest version. This is a brand new HANA in top of non-SAP (DB2 source). What is the best way to model this scenario, meaning, can we model everything in the Calculation view’s like SAP HANA live or do you suggest using the standard attribute, analytical and calculation views to do the data model. Is SAP moving away from AV & AT to only calculation Views to simply the modeling approach?
>> I haven't read any "official" guidance to move away from typical modeling approach, so I'd say stick with the usual approach- AT, then AV, then CA views. I was told that the reason for different approach with HANA Live was to simplify development for mass production of solutions.
Reporting: We are using the design studio as front end tool. Just for example, if we assume that we are
Using the BW, we bring all the data in to BW from different sources, build the cubes and use the bex query. Here in bex query we will be using the restricted key figures, calculated key figures calculations etc. From the reporting wise, we have the same requirements, calculations, RKF, CKF,Sum, Avg etc. if we are Using the design studio on top of standalone HANA, where do I need to implement all these calculations? Is it in different views? (From reporting perspective, if it’s BW system, I would have done all the calculations in BEx.)
>> I'm not a BW guy, but from a HANA perspective - implement them where they make the most sense. In some cases, this is obvious - restricted columns are only available in Analytic Views. Hard to provide more complex advice here - it depends on your scenario(s). Review your training materials, review SCN posts and you should start to develop a better idea of where to model particular requirements. (Most of the time in typical BI scenarios, requirements map nicely to straightforward modeling approaches such as Attribute/Analytic/Calculations Views. However, some situations such as slowly-changing dimensions, certain kinds of calculations (i.e. calc before aggregation with BODS as source - where calculation should be done in ETL logic) etc can be more complex. If you have specific scenarios that you're unsure about, post them here on SCN.
Universe: If we are doing all the calculations in SAP HANA like RKF. CKF and other calculations , what is the point in having additional layer of universe , because the reporting compnets cam access the queries directly on views .In one of our POC , we found that the using universe affect performance.
>>> Depends on what you're doing. Universe generates SQL just like front-end tools, so bad performance implies bad modeling. Generally speaking - universes *can* create more autonomous reporting architecture. But if your scenario doesn't require it - then by all means, avoid the additional layer if there's no added value.
Real time reporting: Our overall objective is to give a real time or close to real time reporting requirements, how data services can help, meaning I can schedule the data loads every 3 or 5 min to pull the data from source. If I am using the Data services, how soon I can get the data in HANA, I know it depends on the no of records and the transformations in between the systems & network speed. Assuming that I will schele the job every 2 min and it will take another 5 min to process the Data services job , is it fair to say the my information will be available on the BOBJ tools with in 10 min from the creation of the records.
Are there any new ETL capabilities included in SP7, I see some additional features included in SP7. Is some of the concepts discussed are still valid, because in SP7 we have star join concept.
>>> Not exactly sure what your question here is. Your limits on BODS are the same as with any other target system - doesn't depend on HANA. The second the record(s) are committed to HANA, they are available. They may be in delta storage, but they're available. You just need to work out how often to schedule BODS - and if your jobs are taking 5 minutes to run, but you're scheduling executions every 2 minutes, you're going to run into problems...
Thanks
Magge
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Thank you Jody for the answers . Indeed it was helpful. On the similar lines , we are trying to use
the Sybase-IQ to store for addtional years of data.
I haeard that , If i use the Sybase-IQ , then i can not use the Attribute or analytic views becasue
i will be doing the join operation on the data on hana and th edata on Sybase IQ.
This means that i need to combine the base tables in the Calculation view ( Both the cases in HANA Plus Sybase IQ) . In both th ecases i can not use the Attribute or analytical views.?
if i use the Sybase IQ-i know we can use the virtaul tables or New extended table. Which one is the
latest , Virtual table or extended table. I read that we can use both to pull the data from Syabse-IQ virtually.?
If i build Calculation views for acountry and if i wanted to implement/model for another country ,
do i need to create another set of views . How do i implement this scenario. Because in BW World
you have one infocube and load the data from multiple countries.?
Thank you
Magge
Hi Magge,
Take a look at SP8 for smart data access.
Support for Analytic & Attribute Views
• Virtual Tables can now be included along with
Physical tables when creating an Analytic &
Attribute Views
• This is don’t use a new ‘migration wizard’ using
the “Mass Copy” feature from HANA Studio.
• When adding a Virtual Table to an
Analytic/Attribute View these are automatically
migrated to Calc Views
Hopefully this would answer your question for sybase IQ.
Regards
PK
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
User | Count |
---|---|
86 | |
10 | |
10 | |
9 | |
6 | |
6 | |
6 | |
5 | |
4 | |
3 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.