Your code fairs no better, Joe.  Try running it and see the error.

Also, the columns in your CREATE TABLE statement don't match the columns in your SELECT statement.

And, your code...

If you avoid "batching", it can be done with "minimal logging", which also nearly doubles the performance.

The title of this thread threw me... I thought we were trying to count the NULLs in a given column... which is different than the original script, as well.

Heh... you didn't try the code you posted.  It's missing commas and the VALUES are not correct.

Be advised that the data you posted has nothing  other than partid and serial...

I believe you'll find that and INSERT/SELECT from an OPENQUERY across a linked server will work with some pretty good performance.

Wouldn't it be easier just to take the difference between the count of a column that will never be NULL (like the PK column) and the column in question?

The typical reason why there are typically no standards, particularly when it concerns databases and the related code, is because of the "code first" mentality that I've seen at oh...

There are a ton of decent articles out there on this subject.  You really need to do a little "googlin'". ...

This is what fixed us a long time ago.  I just...

Yeah... there's always something when it comes to stats.  Like you said, ascending date columns can be quite the rub, as well.  We have a whole lot of tables that...

My original thought in response to this article is "It Depends on the 'tool' in the chair!" 😀

As an infrastructure guy and former consultant, one universal truth that I've observed over the years is the tendency for useful tools that are controlled locally to be replaced...

Don't give the privs to the user.  Instead, write a stored procedure that uses EXECUTE AS OWNER and give the user privs to execute that stored procedure.

In all seriousness, we've always worked in a "distributed fashion".  And, yeah, databases are necessarily a bit different simply because of their size and the fact that they contain sensitive...

I just reran the code I posted on my production box and it's returning precisely (except I stripped the time off the "Last Accessed Date" output, which you can get...