If you change the last line to "group by G.GroupCode WITH ROLLUP", you'll get an extra row with the grand total for Amount (and a NULL for GroupCode).  Is...

How about the ErrorMessage field of the LitespeedActivity table?

NOLOCK is like a chainsaw, it's very useful but also dangerous.  I don't think I would put it on any list of standards or best practices if those recommendations would...

If join performance is a concern for a normalized design, there are a lot of things to check before going to a denormalized structure.

Do the indexes allow efficient joins?  Does...

This will produce the queries to find all occurrences of 312564777 in int or bigint columns, or [n][var]char columns of length 9 or greater, in all user tables with only...

Reading the backup history is great, but it might not show you when backups failed.  For that you need to look at the SQL Agent log.  This is not a...

exec @return = Proc arg1, ...

The only performance gain achieved by using "SELECT *" is that it takes less time to type.  You might argue that there is some miniscule benefit in compilation time, or...

Whenever I see something create a temp table and then delete unwanted rows, I wonder if the final table could be generated in one step.  The ROW_NUMBER function can't be...

How are we supposed to know when we're supposed to spot the sneaky little syntax errors in the trick questions, and when we're supposed to overlook the stupid errors in...

Most of the suggestions in this thread have assumed that you have some control over how the backups are done.  If Netbackup is running things, you don't have any control. ...

One of the compressed-backup vendors (I don't recall which, probably either LiteSpeed or RedGate) has a benchmark of backing up 1TB in < 2hr.  They probably were using hardware that...

A dual-core CPU may be equivalent to 1.6-1.8 single CPU's, but that doesn't mean it is equivalent to 0.8-0.9 x the performance of a 2-CPU system.  It should perform better than a dual-processor...