Question

Are column store tables now the vendor default method for Kimball Data Warehouse tables with Dim and Facts for On Premise?

I know in the cloud, Microsoft automatically recommends Azure Data warehouses use column store. How about on premise? Is this the vendor solution, or should I conduct row based testing?

We have SQL 2016, 50 cores, 200GB of RAM. Database consumes 500GB each year.

Was it helpful?

Solution

As of SQL 2016 Clustered Columnstore Indexes should be your default choice for Fact tables in a Kimball-style dimensional model, possibly with the addition of a non-clustered primary key to prevent duplicate rows. See Columnstore indexes - Design guidance

Dimension tables should typically use row-based Clustered Indexes, with additional non-clustered indexes for business keys.

Licensed under: CC-BY-SA with attribution
Not affiliated with dba.stackexchange
scroll top