Tuesday, 9 July 2013

Clustered Column store index recap

Last week I blogged on the Clustered Column store index, I’ve been playing with it – its very early days, as its a feature of SQL2014 CTP1.

Just to recap.  I noticed that something wasn’t working as I expected.


The number of rows returned from the report ‘Disk usage by table’ is incorrect, shown by the top results set, the correct number is shown from select count (*) query and the select from sys.partitions.

Also notice on sys.partitions, the compression description.


So how do I get the standard report to show the right data?

Well, restarting the instance doesn’t help.

But what about rebuilding / re-organising the index? Re-organising the index does nothing, but rebuilding the index works!


All the counts are returning the expected number of rows, now everything is alright with the world.