Hi everyone,
I faced a strange use case while analyzing SDA consumption with Visualize plan. The data model concept is the following: I have a virtual table under my HANA schema coming from a DB2 database via SDA remote source connection. This tables after filtering in projection contains 2,7 million records and this projection is being aggregated with a use of aggregation node as shown in the screenshot:
I did a two step analysis in Visualize plan: When I just execute the Projection I need around 11 mins to get the 2,7 million records as in the screenshot:
However when I do the same in the Aggregation node in the next step of my model I get the aggregated data in 2 mins as in the screenshot:
To be honest I dont understand the second result. I mean I would expect to get the 2 million records in HANA and get the aggregation there, but instead it seems that I get the aggregated data directly.
And the last part is that when I evaluate the whole model after the aggregation, I have the behavior which I was expecting, meaning I get the 2,7 million records from the remote source and then the aggregation takes place and the whole model last 6-7 mins:
Could you please help me on that to understand how the data are being consumed and why these differences in the time response?