That's parallelism waits, not actual blocking

Have you looked at the execution plans?

Could you please expand the sample data to show when an ID will return more than 1 for a year?

The wait durations in the screenshot are all around 15ms, so probably not significant (timeout is 30 seconds). They're also from all different processes and you haven't indicated which process...

Taking myself and a good book to an expensive Chinese restaurant tonight. Yes, that's a celebration.

I don't know how things are done in the US, but here 75% is the threshold for a degree to be awarded "With Distinction"

No, because like the buffer pool it's not per-process. You should be able to use the memory clerks or other memory DMV to identify what area is using a lot...

Don't try reading the XML plan. Use the graphical one.

You won't find CPU time in the plan. You'll find costs, but they aren't times and can't be used to do...

No. Absolutely not.

There is a Linked server openrowset query...

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

I'm going to assume there's a mistake here, as a single column can't be equal to two values at the same time.

Left-to-right to see the control flow.

Right-to-left to see the data flow.

Careful,...

Union all if there's no overlap or if duplicates are OK. Union if you want to eliminate duplicates.

Queries don't use much memory. The SQLBufferPool clerk is the data cache, the plan cache and a few other caches. It should be the highest memory consumer as that's the...

There are ways to do it, but they're complex, harder to read and usually perform badly, hence why I'm recommending you use two statements. Production code isn't where you show...