Please scroll to the bottom for the update
About three months ago I wrote a blog on “Tableau on HANA”. Much to my surprise I see an interesting comment showing up by Mr. Francois Ajenstat:
Hi Ronald - the SAP HANA connector for Tableau has been released as part of the 7.0.9 maintenance release. You'll have a better connection experience, functionality and performance with HANA.
Now I do not know Francois, but after a quick Google search I find out that Francois is the Director of Product Management for Tableau. Now he might be a bit biased, but still, a new native connection is interesting enough to investigate. And so I find myself once more behind my trusty Mac
It always starts with a download (again)
Indeed, a quick download is required again. You need to have the latest version of Tableau, so be sure to pick it up here.
Connection to a table
Now instead of connecting to my “sales” table, let’s try something different and connect to my “Twitter” table, which I used in my last blog: The not so fuzzy fuzzy search.
Connection is still a breeze. Simply choose “SAP HANA”, fill in your server, username and password and select your schema:
Tableau will ask you how to connect to your data. Be sure to select “live data”. One interesting thing is that Tableau does not come with warning messages as with the ODBC connection. Seems we’re good to go!
Time to build a report
Now, it’s still not a very smart idea to try to analyze 50 million records in one go. Be smart with your selections and do not forget to pause “auto update” or you will still be finding yourself drinking several cups of coffee before the result set comes back.
I create a simple report. A horizontal bar chart with per month showing count of Tweets (based on the contents of the Tweets). I create a quick filter on the content and select “Wildcard Match” in order to be able to do a broad search on the tweets I loaded.
The result looks something like this:
Note that I selected “apple” in my filter. The result comes back without any delay. The great thing is that I can actually take a look at the underlying data by selecting “view data”. Over 12000 records are found containing the word “Apple”. You see some of the top results here:
Time to build another report
Now what happens when I build the same report, but use the ODBC connection to HANA. Well actually not a lot in terms of speed, data still comes back in a second. What is more interesting is the results I get back from the query:
Completely different result set! The search comes back with all records where the letters in the word “apple” match. Even though this comes close to a fuzzy search ;-), it’s safe to say that the new connection is much more accurate than the ODBC one for this example.
Time to build another other report
Alright then, one more. This time we build our report on an analysis view and consume that view in HANA. According to Tableau that is supported now.
Please note that views are not in your default schema, but in the “_SYS_BIC” schema. My connection would look something like this:
Note that the package is part of the table name.
I drag my fields in the rows and columns and…
No sigar…
Now I believe this a general limitation in the HANA view (it is even an error message coming back from the generated SQL).
Seems the SQL statement is missing a "GROUP BY". Putting the SQL statement in the HANA studio also generates the error:
Update 14-10-2013
As rajiv.bahl
Please leave your comments below!
Take care,
Ronald
- SAP Managed Tags:
4 Comments
