cancel
Showing results for 
Search instead for 
Did you mean: 

How to define measure in HANA Studio for Lumira Box Plot

former_member190800
Contributor
0 Kudos

Hello Colleagues & Community,

I have a data column in HANA DB related to duration.  For example, I have instrumented a mobile library, and am recording the results of various timers, together with information on the current conditions.

Since I am interested in a statistical view of the values, I would like to use a Box Plot in SAP Lumira (MIN, Q1, MED, Q3, MAX).  I am exposing the data to Lumira via an Analytic View created in SAP HANA Studio.  The Analytic View logically joins a data foundation containing the fact and dimension tables, with a time-series attribute view.

The issue or question that I am running into is:  How can I define a measure on the duration column in the Analytic View, without specifying an Aggregation (COUNT, SUM, MIN, MAX) type?


The reason I think I want this, is that Lumira should be able to access the raw data points of the duration column, in order to calculate the Box-Plot quartiles.  Instead, I'm seeing it attempt to calculate quartiles on the SUM of the duration column over the period, and it's just returning the SUM for all quartile values--a horizontal line.


Any ideas?

Thanks!! Stan


Screenshots:

Here is the raw data in HANA Studio


Here's the Semantic definition of the Analytic View (TIMERDURATION has Aggregation options: SUM, MIN, MAX, COUNT)

And, here's the Box Plot in Lumira (note that both values are the sum of the raw data above)



Accepted Solutions (1)

Accepted Solutions (1)

Henry_Banks
Product and Topic Expert
Product and Topic Expert
0 Kudos

Hi,

in lumira desktop, are you able to go to the Measure > Change aggregation > and use None ?

regards,

H

former_member190800
Contributor
0 Kudos

Hi that looks exactly right--for created measures-- and I can see the documentation here:  Creating a measure from a column or a dimension - About SAP Lumira - SAP Library

However, the "Create a measure" option is disabled on the target column.  It's a numeric type... I see a note that I must be in Facets mode, and I think I am...  does this look familiar?

Thanks,

Stan

Former Member
0 Kudos

Hi Stan,

Could you create a Calculated Column (of type Measure) on in the calculation view, which wraps around Duration measure and set the aggregation to None (the other option is Formula) for this newly created column?

Regards,

Boris

former_member190800
Contributor
0 Kudos

do you mean:  create a Calculation View in HANA Studio?  Or, in Lumira Desktop Client, create a calculated column on top of the measure?

Former Member
0 Kudos

In HANA Studio.

former_member190800
Contributor
0 Kudos

I found the location I think you were describing, but here also the "Change Aggregation" option is disabled.

I wonder if this is related to using "Connect to SAP HANA" as a data source?

former_member190800
Contributor
0 Kudos

I found that if I download the data from SAP HANA for the data source, I can set aggregation type to "None".  !  I'm not sure if that's by design, or a bug--do you know why this would be?

So, I will try Boris's recommendation also.

Thanks,

Stan!

Former Member
0 Kudos

Hi Stan,

You are correct. The change aggregation is disabled for online HANA views, and that's by design. My suggestion above will not work either.

Regards,

Boris

former_member190800
Contributor
0 Kudos

Thanks for the clarification.  Why is that? 

Former Member
0 Kudos

For online HANA views, aggregation is done natively on HANA server side, which itself does not support None as an aggregation type. This is also a restriction when publishing an offline dataset to HANA containing a measure with Aggregation type None (page 129, http://help.sap.com/businessobject/product_guides/vi01/en/lum_122_user_en.pdf).

Henry_Banks
Product and Topic Expert
Product and Topic Expert
0 Kudos

Hey Stan, 

in future, as an SAP employee please could you bring your Lumira questions to the internal JAM group (BIPedia)?  - we try to keep SCN for customers.

Many thanks,

H

Answers (0)