SQL Server: Grouped Aggregate Pushdown. Make columnstore indexes go faster with aggregate pushdown functionality!
Hi Guys, Today as promised we will talk about an important feauture on the columnstore indexes called Grouped Aggregate Pushdown . We already talked about this feature two post ago here: SQL Server, the "Group By" competition: from 30 seconds to 56 milliseconds and here: SQL Server, the "Distinct clause" competition. Oops i did it again! ...from 24 seconds to 6 milliseconds! We have seen that a columnstore index can return directly already filtered data so this feature is very powerful ! I want to repeat the official definition because in my opinion it is very beautiful: Some aggregate computations can be pushed to the scan node instead of calculated in an aggregate node. Where is the problem? Activating this function is not really trivial! Let's see, even with the help of a few examples, how to do it. Enjoy the reading mate! "Summary of previous episodes".. let's recap! In the example we had done in the previous post we had run the followi...