I have heard this from a couple of sources but never found any solid documentation to support it. Does the restore process contain some mechanism to update statistics? On the one hand restore will upgrade a database so it seems reasonable from that standpoint. But if it did update stats you would think Microsoft would add a parameter to turn it on or off. Add to that experts like Paul Randall repeatedly state that a backup is an exact copy of the database at the point of backup and this starts to smell like a myth. So let’s see if it is true or not.
Let’s look at the statistics of AdventureWorks2008 before taking any action. To narrow things down let’s look at the DatabaseLog table in particular.
Now let’s update the statistics by running sp_updatestats and take another look.
As you can see the statistics have been updated for the primary key on that table. Now let’s restore from a backup made prior to updating the
statistics and see what happens.
The statistics post restore have returned to the original state. Based on this experiment this can be put in the myth category.