Star Joins in today's HANA

[shyam_uthaman]

Hi,

I've asked this question a long while ago but did not get a satisfactory response.

So.. here goes another try..

As we know, there is a star join option while creating a calculation view graphically.

I've been hearing from certain people that creating a view using the star join is better for performance.

I've never found evidence to support this but would like a healthy discussion on the below questions:

1. Does it indeed offer any performance benefits?

2. If yes, under what scenarios would it be used ? I've lived perfectly fine not using them but are there special cases when they perform better? If yes, an example scenario might be great for my understanding.

Thanks in advance.

Shyam Uthaman

[lbreddemann]

STAR JOINS are the functional equivalent of analytic views.

The idea here is to use this modelling artefact so that HANA understands that what you are working with is indeed modelled like a star schema.
Also, analytic views and views with a star join are expected to be queried from in a specific way:

filtering on the dimension attributes and aggregating measures/key figures from the fact table.

For this scenario (which mostly occurs in data warehouses but not so much in master-detail OLTP scenarios) the star join does the job better.

[former_member228588]

Hi Shyam,

Let me tell you are scenario where i have used the Star Join, and without Star Join, i was not able to achieve my desired result.

You can see that in my blog :-

https://blogs.sap.com/2017/01/26/advance-hierarchy-modeling-with-sap-hana/

Hope this will help you.

Thanks