Yes, I can use the check constraint for the field. But I need the additional constraint of uniqueness for specific types of data, while not for others. I don't think...

Looks great. Many thanks to both you and Mike - I had no idea this feature even existed (SQL Server is kind of overwhelming sometimes). This takes care of my...

Thank you, Steve. I just looked at some intro tutorials. It doesn't seem as bad as I though at first. I enabled Filestream in the DB, according to instructions on...

Hm, okay - that's an interesting idea.

This was originally an Access database (actually, it was initially a series of Excel spreadsheets, before I got involved), so packing this much attachment...

1. This does the same thing as the PowerShell command, and runs at about the same speed.
2. I don't know what you mean here. I thought filestream was for reading...

Look at what the resulting values are that come out of the hash match - in the first query the number of rows is less than what you see...

Looking at the difference between plans:

The first plan utilizes an index scan using the index ix_TaxLok with no predicate.  This results in an estimated row count of 124720 and...

Nope - nothing changed. The plan looks the same, the estimates are the same, and it even...

• This reply was modified 5 years, 8 months ago by pdanes.

I found this in the second plan and not in the first. Interestingly, the second plan...

Think of "Left Join" as a different operator from "Inner Join".  That is why SQL Server is needing to generate a different plan.  Unless you put in parentheses to...

Okay, that didn't take. Here is a zipped version.

Here they are.

I also swapped the order of the two queries and ran them again. I both cases, the query first in execution order took more time than the second. Here is...

SQL Server parse...

I think the order of operations regarding the left join is coming into play here.  The optimizer is leaving the door open to the join between P and T...