Hi Ravindra,
I have 2 models whose output is same but differed by design.
In model 1 Projection_1 acts as the input for Projection_2 and Join_12. In Projection_2 half of the columns are used and the data used is half the amount present in Projection_1 .
In model 2 we are using 2 projections for the inputs , Projection_1 serves as the input for Join_12 and input to join_2 is replaced by a
single projection with a filter and selecting only those columns which we require for further proccesing.Remaining design remains the same for both the
models.
Join_12 is left outer join where Projection_1 is the left element and projection_12 is the right element.
When we execute both the models we find that the model 1 is faster than model 2 , even though Projection_2 in model 1 queries the entire set of data and then applies filter to it. And in model 2 the data fetched from db is a filtered set with only required columns.
Looking for your inputs.
Regards
Rohit