Hi,
As with any unbalanced compare, the performance varies.
Just compare this method with use your favorite compare and look closely at the execution plan
In my query close to 50% of the time is spent reading the 2 source tables using a clustered index scan, which is the most efficient way to read the ENTIRE table content.
The Merge join that does the actual compare work is only 19 % of the time
The rest: 33%, is spent storing the result.
If you find a more efficient way, i'd be happy to review your code!
Theo 🙂