Me too, have been running this on SQL 2008R2 - 2014.

Only issue I have seen is if a legacy procedure has been doing a mondo-sized select-into.

The execution just hangs. ...

This is most curious.

I have been using this for quite a while.

What MS SQL Server version are you using?

You are welcome.

David McKinney (3/25/2014)

---

you can use UNION also to do something similar. In this case, it returns both rows when there's a difference.

SELECT*

FROM(SELECT*

,count(1) OVER (PARTITION BY ID) AS versions

...

thank you for your explanation.

having rows in the source that do not exist in the target as you describe is precisely why the join exists and the test...

Without the join, the except statement would also return rows in the source that do not exist in the target as well as those that exist (by id) in both...

Where I have been using this is to keep read-only data synchronized between two areas by their primary key, so I do not currently have a scenario like this. ...