With thanks to the ASUG Virginia Chapter volunteers, particularly Keith Seckman, Rich Heilman Thomas Jung and sponsor DataXstream we just finished a two day hands-on workshop “SAP HANA Developers”. While I am not a developer, I learned a great deal, with some of my usual notes below.
Figure 1: Source: SAP
All new development will be on ABAP for Eclipse, not SE80
What need to do on your code to run on HANA
With BW on HANA you don’t have so much custom development to leverage HANA
You can have new applications; it is fast, in memory, column store, no indexes
Thomas said it was more than a fast database
It also has text analysis, spatial, predictive analytics
HANA native allows you to run directly out of database itself
Workshop – new apps and business suite
Figure 2: Source: SAP
With on premise you can do custom development with ABAP or HANA native
Speed development – ABAP or HANA native
2 different – private managed cloud HEC or HANA Enterprise Cloud where you are paying someone to manage hardware and software for you.
You can’t do Java custom development in HEC – only ABAP and HANA native
With the public cloud you can run on another database; the infrastructure is multi-tenant environment – no ABAP development
Figure 3: Source: SAP
Figure 3 shows that SAP HANA is a “complete relational database” Join, in memory columnar – same relational SQL
Migration is a technical part for BASIS using the DBTOOl, unload, reload, need to do an upgrade to 740 (plus Unicode)
Performance is where you spend time as your company expects better performance – if report SELECT * FROM VBAP for each record in header table – HANA can speed up a little more
“Do things to let HANA help you” – use WHERE conditions
Code pushdown – put as much of the processing in database – not do so much inside internal tables
“Write better ABAP code” was the statement made during the workshop.
Figure 4: Source: SAP
Figure 4 shows “What happens during migration”
Figure 5: Source: SAP
SAP owns database and ABAP kernel so it can optimize things to interface better
FDA “fast data access” to transfer data more efficiently
Rule is still the same, minimize transfer
Instead of a million select single, use an INNERJOIN and let database do the work
Minimize search overhead – no secondary indices with HANA any more
Store data as columns and not as rows
Put work into HANA to see benefit
Figure 6: Source: SAP
Figure 6 shows pushing the code to database to take advantage of database
You shrink the app server
Let database take the load
More to come as I have time.
One of the final exercises was creating a Fiori launchpad – success! The other nice part is we could download our work and then use it on a SAP CAL/Amazon instance.
The next ASUG Virginia Chapter meeting is October 3rd – I believe in Williamsburg, Virginia, co-located with a golf outing.