Pages

Wednesday, 26 June 2013

Interesting thing about updateable Clustered Column Store Index

I downloaded SQL Server 2014 CPT1 last night and I’ve started playing with it.

I’m really interested in the updateable clustered column store index, so I thought I would see if it works in the CTP1 (not all features are released in the CTPs)

I created two identical tables, added a Clustered Column store index to one of them, and inserted 15000 rows into both of them.

Guess what, it only bloody worked! Excellent! Happy bunny.

So the next step, how does it work? While looking at the IO for the insert, I stumbled on this..

image

TestTB_1 has 15,000 rows and TestTB_2 has 0 rows… Wait a second, that’s not right.

image

No – There are 15,000 records in both tables.

image 

Wow – check out the logical reads,physical reads…etc…etc..

All coming back as 0 – I guess the column store engine isnt reporting back in the same way the normal engine is.

This will get fixed at some point, no doubt, but the fact that the column store index was updated is so cool…

I’m going to do some more digging now…