First things first...what's "R"? Simply a programming language and software environment for statistical computing and graphics. More infomation can be found here R on Wikipedia


I have code in many programming languages, some of them very commercial, and some of them little known, but I got say, that from all, "R" is one of the most weird and awesome languages I have ever played with...and it has an amazing repository of custom add-ons.


If you have read the HANA Pocketbook you will realize that there's a reference to "R" in the page 59. Now, that kind of integration between "R" and HANA haven't been developed yet, but it doesn't mean we can get our hands dirty doing some research and development.


What I did for this example was to simply show the information of my Analytic View on HANA and exported as an CSV file. From there, it's easy to import it into "R" and start doing some nice things. (The idea is that we should be able to code "R" straight in the HANA environment...or at least that's how I think it's going to be...)








The first example that we're going to build on "R" is a simple Pie graphic, using the information from the FORCURAM and CARRNAME fileds.




In this example, we're basically reading the CSV file, including the header. And doing an aggregation of the two fields we want to interact with. After that, it's just a matter of pass the values, the names and call the pie.




Next example is a little bit more complex...and uses a custom package call Word Clouds




Here, we have to load the required libraries, read the CSV file, do the aggregation, create a matrix with the aggregation values, sort the matrix, create a new vector, get it's length, create an array containing the names and finally assign the values and call the wordcloud graphic method...




Hope you like it...and stay tuned for more "R"...

Lars Breddemann

HANA backtrace 45/2011

Posted by Lars Breddemann Nov 16, 2011

Ok, it's been more than the planned week since my last post in this series (HANA backtrace 44/2011) but these are very busy times lately for a primary support guy like I am.

I write this post from our US Headquarter offices in Newtown Square near Philadelphia and that's because I've been asked to deliver several internal HANA training workshops for our AGS colleagues over here. Thus, I booked the flight, packed my bag and flew over here.
And there we are - right in the middle of a quite packed training week, covering topics like performance analysis, modelling, data services and backup/recovery (and lots more).
And just as we're in full flight, we see HANA SPS 03 being released to customers.

With this, many new features, usage types and bug fixes have been made available to HANA users.
Make sure not to miss out on the updated SAP HANA documentation

For the HANA database this is revision 20 and the changes that have been made in this revision are documented in SAP note #1648901 SAP HANA database Revision 20.

The actual SPS 03 release notes can be found as an attachment to SAP note #1642937 SAP HANA Appliance Software SPS 03 Release Note.

++++ ---- ++++ ---- ++++

With SPS 03 a new software installation tool (Software Update Manager, SUM)  had been released for HANA systems.
See SAP note #1545815 SAP Release Note for SUM for SAP HANA 1.0 to get the details.
The nice thing with this is, you can easily manage the upgrades of HANA related software parts all at once.

Also very important to know is that with SPS 3 now log backups are supported with the HANA database.
Note #1645183 Changing log mode for SAP HANA DB after SPS2 to SPS3 upgrade talks about the changes in log mode that have to be implemented for this to get activated.
++++ ---- ++++ ---- ++++

With SPS 03 a new way to access data from a HANA system had been released to customer: the web-based SAP HANA Information Composer. Find details in the following two notes

#1627904 SAP HANA Information Composer: Central Note
#1633168 SAP HANA Information Composer: SP3 Release Note
++++ ---- ++++ ---- ++++

Just a small code fix for a data type change for function module TREX_EXT_SHOW_INDEX is delivered by sap note
#1652229 SAP HANA: error code -10404 with TREX_EXT_SHOW_INDEX
++++ ---- ++++ ---- ++++

As far as I know the pricing in HANA is based on the amount of data used with HANA, so the importance of a proper sizing is self-explanatory. SAP Note
#1637145 SAP BW on HANA: Sizing SAP In-Memory Database
contains the information about how size a HANA system if you want to migrate your existing SAP BW system to it.

Which reminds me: BW on HANA has also been made available by the release of HANA SPS 03 and SAP BW SP5.
++++ ---- ++++ ---- ++++

Also related to the migration of ABAP systems to run on HANA is note
#1634681 DB Migration: Report for finding large Row Store Tables
which basically makes sure that you don't overlook very large tables that should be put to the row store.
Why can this be an issue?
That's because unlike column store tables, row store tables are always fully loaded into memory on HANA instance startup. So if too much data is put into such tables, it might be that the instance won't come up any more, since the automatic loading of these tables fails.
++++ ---- ++++ ---- ++++

As with all NetWeaver main databases, which is a use case for HANA as well (e.g. BW on HANA) there is the usage of a database sequence called DDLOG_SEQ. Unfortunately older HANA revisions didn't support persisted sequences, so after a restart of the HANA instance the instance would actually be reset to its initial value.
To prevent this, see SAP note #1648526 Incorrect values for DDLOG sequence.

Talking about HANA usage with NetWeaver systems, the database interface library (DBSL) is a core piece of technology - so make sure to have SAP note #1600066 - verfügbare Patches der DBSL für NewDB on your radar to keep track of the latest fixes to it.

And, while you're already at it, just also check SAP note #1597627 - HANA connection about the secondary database connection with HANA.

In case you should check your SM21 system log from time to time and find error messages that indicate 'database system is not supported' since you're running on HANA, then SAP note #1650209 SM21 Meldung: "Datenbank System wird nicht unterstützt" is for you.

++++ ---- ++++ ---- ++++

I was quite happy to learn that there are also help tools for HANA created for mere support folks like me when I found SAP note #1647717 Für den SAP-Support: Prüfprogramm für Tabellen.
This note contains a little check program that looks out for tables stored in HANAs column store that might require a MERGE of the delta indexes (in case the automatic MERGE had been turned off).

Nicely also SAP note #1652078 SAP HANA database: Hardware check is also kinda support related.
Since HANA revision 18, a little python script is delivered with HANA and executed during installation time.
It basically checks and reports whether your hardware is actually supported for HANA and it looks like my test instance definitively is not... :(

logon as root...
vml3012:/usr/sap/HAN/SYS/exe/hdb/python_support # python
performing Hardware check......
could not get Hardware Information: 'CPU Flags'
assuming largest storage device  /sapmnt/vml3012/a as data device....
assuming 2nd largest storage device  /dev/shm as log device....
System Manufacturer            VMware, Inc.              not supported
System Product Name            VMware Virtual Platform   not supported
could not get Hardware Information: 'CPU Model'
Hardware check failed!!

As a final support related sap note there is #1641210  SAP HANA datebase: Checking for suspected problems.
So, in case there are any kind of issues, it's sure a good idea to go through the steps mentioned in this one to avoid the common "traps".
++++ ---- ++++ ---- ++++

For the SLT users, the following SAP notes are surely a must read/implement:

#1605140 SAP HANA 1.0: Central Note - SAP LT Replication Server
#1647648 HANA LTR - corrections for DMIS 2010 SP4 111031 (stop repl)
#1625755 SAP LT Replicator - individual table settings
#1649910 Applying SLT (DMIS_2010 SP05)with SAP HANA SPS03
#1650191 Syntax change in NEW-DB 111108
++++ ---- ++++ ---- ++++

In case you're running HANA on IBM storage you should give SAP note #1650046 GPFS maintenance after SAP HANA node has failed a read to learn about the required checks to be performed when a cluster node failover happened.
++++ ---- ++++ ---- ++++

Other releases

As I barely had time to go through the notes while preparing my training sessions, I didn't make it to the blogs on HANA this week. However, I've seen that there had been several HANA related posting, especially in connection to the recent SAP TechED, so give yourself a treat and check "SAP Blogs on In-Memory Business Data Management"  to catch up with them

So, that's it for this week. I expect more to come for HANA, so see you next week.

This week I'm starting a new blog post series aimed at the readers that are technically interested in SAP HANA but don't want to spend hours each week checking every new post, article, forum discussion or sap note about this hot topic.

These regular posts will be little collections of what I found interesting, notable or important in the area of HANA.
Since I'm rather the technology guy and severely biased towards the database-backend-part of it, it shouldn't surprise anybody that this is not going to be a comprehensive overview that covers all HANA related topics.
Instead it'll merely be what I read on the topic - so I hope you like this :-)

If one or the other now thinks "hmm... this concept is somewhat familiar..." then you're absolutely right. I really like Pythians "Log Buffer") a lot and HANA is just missing there.

So let's start and look what the last week (CW 44, 29.10. - 5.11.2011) had for the HANA community:

Several SAP notes had been published

The most important to start with: Revision 18 of the SAP HANA database (and of the related client software of course) had been released and it fixes a lot of issues. On top of that, it also brings several enhancements, especially in the area of the error handling with delta index merging... at least this is mentioned twice in SAP note #1646454 SAP HANA appliance: Revision 18 of SAP HANA database

o Improved error handling for delta merge
o Improved delta merge error handling

Maybe they double-enhanced it ... :-)

++++ ---- ++++ ---- ++++

If you're using the segment builder from the SAP CRM solution and want to access HANA then SAP note #1577019 SAP HANA Database: Abap module for Segment Builder is for you.

++++ ---- ++++ ---- ++++

SAP BW on HANA is about to be published soon, so sizing is a topic that will become important. Details on this can be found in note #1637145 SAP BW on HANA: Sizing SAP In-Memory Database.

++++ ---- ++++ ---- ++++

HANA instances can of course be monitored via the DBACockpit like any other supported DBMS. As a security enhancement there is now a special user (DBACOCKPIT<SID>) setup available that allows the DBACockpit usage with mininmal permissions for the monitoring user account.
SAP note #1640741 Different database user for DBA Cockpit for SAP HANA has all the details.

++++ ---- ++++ ---- ++++

If you want to install all already published SAP BASIS corrections in one go, make sure to check SAP note #1644214 Comp. SAP Note Basis porting to SAP HANA DB, NW 7.30 SP06.
It's a composite note that contains all HANA related corrections for SAP BASIS 7.30.

++++ ---- ++++ ---- ++++

In case you're moving your SAP BW system to HANA you won't need to export the aggregate fact tables, since these aren't necessary anymore with HANA. The correction from SAP note #1644396 SMIGR: Kein Datenexport der Aggregate-Tabellen modifies the famous SMIGR_CREATE_DDL report accordingly.

++++ ---- ++++ ---- ++++

For the rare cases where you want to create a database index just for a HANA instance, you need to set this up in the ABAP dictionary. SAP note
#1644788 Datenbank-spez. Index: SAP HANA DB ist nicht in Auswahlliste now allows to actually select HDB in the GUI by adding the value 'HDB' to the list of values of domain DDIXDBSEL.

++++ ---- ++++ ---- ++++

SQL Hints will still be with us with HANA and SAP note #1645317 SQL Hint wird nicht im Datenbank Trace angezeigt makes sure that we'll even see them in the SQL trace (ST05).

++++ ---- ++++ ---- ++++

By using secondary database connections it's possible to run your reports agains several HANA instances. Unfortunately there had been a bug with the initialization of the connection information - but no more: SAP note #1646010 SAP HANA: different DB connections are not handled correctly fixes this!

++++ ---- ++++ ---- ++++

Altough HANA database provides many different client technologies SAP provides support just for a defined set of them.
Make sure to check SAP note #1577128 Supported clients for SAP HANA 1.0 to figure out, which combination of HANA database and clients are actually supported.

++++ ---- ++++ ---- ++++

Finally two SAP notes for the trigger based data replication solution for HANA (LTR for Landscape TRansformation, not for L-language TRigger, or License To Rage ... ) provide fixes for bugs:
#1647648 HANA LTR - corrections for DMIS 2010 SP4 111031 (stop repl) #1648783 HANA LTR - corr for DMIS 2010 SP4 11/11/03 (pool 16+keys)

Other releases

In Siri + HANA Sina Moatamed discusses future use cases of HANA and interface technology like Apples Siri.

++++ ---- ++++ ---- ++++

SAP BW on HANA will surely be the next big thing and a lot of people have a lot of different guessing about this. In How BW-on-HANA Combines With Native HANA Thomas Zurek explains how it's actually going to work. Definitively worth the time to read!

++++ ---- ++++ ---- ++++

In Why SAP HANA 1.0 SP03 - Project Orange - will be a runaway success John Appleby explains why BW on HANA really will turn out great.

++++ ---- ++++ ---- ++++

Test driving HANA is what many customers and partners ask for and Jeff Word lays out the current options in his blog SAP HANA Test Drive in Madrid.

++++ ---- ++++ ---- ++++

If you're now really interested in becoming involved with HANA you may check Paul Desmonds Experienced HANA? How to Educate Yourself about SAP HANA blog post.

Et fini!

That's it for this week and for the first issue of this little back trace of my HANA readings. Hope you liked it.

A nice weekend to everyone!

Sina Moatamed

Siri + HANA

Posted by Sina Moatamed Nov 2, 2011

I was having a discussion with a colleague of mine about the significance of Apple’s Siri.  This is the voice activated digital assistant application that is now part of the iPhone 4S.  Basically you make verbal requests of your phone and it, in turn uses different Internet based services to retrieve that data for you.  So for instance you could ask “when did Obama become president”.  Your iPhone might tell you that “Obama was elected president November 4th, 2008.  But he was sworn in on January 20th, 2009”.


Many are speculating that this application could represent the downfall of Google.  Siri will definitely use Google to search the web.  The challenge is in what’s missing from the above simple example.  The user no longer has any relationship to the Google paid links.  The center and bulk of all revenue generation for Google is being cut out of this new UI experience.  For all those other free applications Google provides like Google Apps and YouTube, we may see a new cost model coming if Siri really takes off.


Now, I won’t speculate on the future of Google, but this technology has the potential to profoundly disrupt.  Recently, mobile device sales have outpaced PCs.  This trend is significant and major technology companies are responding accordingly. Couple that footprint with the UI experience of an actually functioning voice driven interface, and you have the making of something pretty significant.


With in-memory technology we are seeing the potential end of data warehousing.  By itself, in-memory technology is a disruptive advancement, but what about in combination with Siri?  Now, I know SAP has made great investments in mobility applications especially with the Sybase acquisition.  They have created an environment where applications can be development and rolled out to any mobile device.  This reduces the time of development and makes the user experience consistent thanks to HTML5 standardization of the UI.  But wait a minute; aren’t we witnessing a complete disruption of what a mobile interface experience will look like?


Now, I’m not proposing that Sybase was an unwise decision or that people won’t want mobile applications on device. But who wants to touch a bunch of radio buttons or type, when all I want to know is what last year sales of our super widget were in North America.


For those who have experience with SAP’s ByDesign, we have already witnessed probably five different UI experiences in its young life. So don’t stop now.  Soon ByD will be based on HANA technology as are SAP’s other OnDemand platforms. My hope is for development to begin sprinting toward Siri enabled mobility. The combination, I think will bring business data to a new place of relevance, both in analytics and application use.

Real cloud computing in my view is about creating agile enterprises, both technically and culturally.  Whether we are looking at rapid deployment of systems or applications, accelerated application development platforms, quick integration, in-memory data and real-time analytics, or mobile access to data from any location, we are pursuing an in the moment relationship to relevant information.  So the next time, I use a mobile device to access my business data, I hope I can just talk it out.


Filter Blog

By author:
By date:
By tag: