Can HANA be used for broadcasting or like a APD ??

[Former Member]

THe queries are running for long time, the BW performance is not good, its lousy, we have been hearing this from the user for quit sometime now.

The users endup with such statement because they are running huge detailed reports , with about 15 characteristics and 5 key figures and execution is for minimum 2 years..(2009 to 2011).. The data is not less than million records.

So we now have to make a choice for BWA or HANA to fix this problem. BWA or HANA would fix the query execution time.
But , we also have another problem, i.e, this huge detailed report has to be broadcasted or exported to the sharepoint or application server. We tried using the precalculation server, but it was not of much help.
I also understand that we have APD tool to export such huge data, but APD has limitations, like variable selection screen should not have "mandatory" fields.. etc...

Will HANA take care of this issue as well ??..
So can HANA export huge data to application server, do this tool has such a feature ??

[Former Member]

Hi Pac,

transaction RSCRM_BAPI might be usable for extracting lots of data from a BEx-Query. This data can then be send to a file or tabel.

Regards,

Leo

[Former Member]

Hi Pac,

word from one of developers of APD -> it was never intended as tool to export huge data out of BW. For that OpenHUB interface was developed. APD was developed for advanced analytical data massaging (like datamining) primarily inside BW.

Best regards, Richard.

[former_member184458]

Pac,

if considering BWA vs. BW on HANA, then BW on HANA is definitely the way to go. Please have a look at this blog

https://www.experiencesaphana.com/community/blogs/blog/2012/08/24/new-sap-bw

and the new BW on HANA FAQ

Http://spr.ly/bwonhanafaq

With regards to the "report export", or may I say data dump, it might be helpful to look at the Business case, as other contributors already suggested.

If the Business case is related to

1. data consumption as-is, the recommendation is to look at the SAP BusinessObjects Explorer on top of BW on HANA as the ideal tool to explore large data sets.

2. further data manipulation/transformation, you might want to look at

2a: SAP DataServices as an ETL tool for BW if the main aspect is to support a scalable ETL process

2b: SAP HANA data mart on BW on HANA, if the main aspect is to provide a SAP data set to analysts, who are more comfortable writing SQLScripts than using SAP BW tools.

BW on HANA provides the BW 7.3 application layer on top of the HANA DB, as well as the possibility to run a SAP HANA data mart on the same system.

Currently supported is the consumption of the "SQL" data models of the HANA data mart in BW via Transient BW InfoProviders without data replication/redundancy/latency. This data flow also allows the use of the sophisticated BW authorization for these data.

In the future, the consumption of data of BW data models in the HANA data mart will also be supported.

This X(cross)-consumption of both approaches will combine the best of both worlds on a single data set in one HANA data base without the need to reconcile data across system landscapes, what is most likely also a big issue in your current set up.

For more details on this aspect, please also have a look at this blog on ExperienceSAPHANA, which provides information about the SAP EDW Enterprise DataWarehouse portfolio: https://www.experiencesaphana.com/community/blogs/blog/2012/08/24/saps-roadmap-for-our-data-warehous...

thanks

erich

[rama_shankar3]

Pac:

   Depending on your client or company culture / mindset they might still want to run the fancy detailed data dump query and get broadcast done. Believe me; I have been in your shoes / situation several times.

BWA or HANA DB for BW will definitely boost the performance of the query but introduces limitations for broadcasting. I suggest you recommend upgrading the client BW database to BW HANA. BWA is not a viable long-term solution since HANA is going to supersede BWA.

Based on what I read, it appears that you might have lot of calculations within the BEx queries or a bad data model on the whole.

    Here are few things for you to try in the meanwhile:

1-    1- Check to see whether the data model is ok by running a RSRV checks for transaction data

2-    2 -  Check to see whether info cube  is partitioned

3-    3 - Try creating aggregates to boost the performance – also cleanup unused aggregates

Hope this helps.

Regards,

Rama

[former_member184768]

Hi Pac,

You've got quite an interesting situation, I must say. It is a very classic situation and pardon me if I say, an abuse of Data warehousing technology. Rather than using BW system for reporting, it seems that it has been used as a data staging application and data pulled out for further analysis (query output with millions of records to be processed outside BW system).

So in this case, whether you use BWA or HANA, I really doubt if it would help, unless you re-look into the very basic need of having this data (of million records) generated as the output.

But the above is my personal opinion about your system.

Coming to the questions you raised..

- Can BWA help you will million records output : Even on BWA, there is a restriction on no of cells BWA can put in the output. So there are chances that even your BWA query will crash if the no of cells go beyond the predefined limit.

- Can HANA generate such an output: Sure, provided you have sufficient memory to process the data, but I still think there are chances that it might run out of temporary space and may not provide the desired output (depending upon the data volume).

As you mentioned, there are restrictions on APD, so considering the high data volume and some flexibility required, one option might be to write an ABAP program to generate the required output. The output may not be formatted and some of the basic query functionality will have to be coded, but that might provide you the required output as per the volume.

Coming to Native HANA, there are export tools, but it would dump data from the table. Again I am not very sure how will that behave for the huge data volume you are expecting.

Going back to my previous point, with Native HANA, I am sure whatever analysis is being done with the exported data, can be performed in much efficient way. So I'd rather recommend you to look into the business case for generating the huge data dump from BW / HANA system and its subsequent use.

Regards,

Ravi