• mishka-723908 (10/21/2016)


    We have a DataWarehouse with Large Fact Tables . The Some of the tables have about 30 to 60 Columns and we also have some denormalized tables that are nearing 300 columns. We are using SQL Server 2014

    We would like to test and see if ColumnStore indexing would make sense for us.

    We load several million of rows and delete several million of rows a day. We keep all history.

    I am wondering if anyone has experience with ColumnStore especially when it comes to deleting large amounts of data daily.

    Thank you in advance

    Couple questions:

    1. Clustered or non-clustered?

    2. When/how are you deleting millions of rows a day? As part of a regularly scheduled ETL job?

    "I cant stress enough the importance of switching from a sequential files mindset to set-based thinking. After you make the switch, you can spend your time tuning and optimizing your queries instead of maintaining lengthy, poor-performing code."

    -- Itzik Ben-Gan 2001