How can I watch overhead when pasting a uniqueidentifier into a primary key column?

As has been discussed many times and is a well-known fact - the presence of a clustered index in the uniqueidentifier primary key column in the sql server will affect performance.

What I would like to do is to observe this problem by releasing inserts and measuring certain indicators. I have 2 tables with a unique PK identifier; one with "newid ()" by default, and the other with "newsequentialid ()" as the default value. I plan to insert about a million rows in each and monitor INSERT performance.

What should I request, in particular? How can I best observe performance gains when using sequential GUIDs at random GUIDs? Are there certain system views, stored procedures that provide statistics about index page splits and / or other relevant information?

+5
source share
2 answers

Although the link below will give a comparison between int and guid, but you can follow a similar approach and try. https://dba.stackexchange.com/questions/9821/why-do-sequential-guid-keys-perform-faster-than-sequential-int-keys-in-my-test-c

+1
source

, . , : 95% .

0

All Articles