Edited it. Not sure where my brain was.

patrickmcginnis59 10839 (10/22/2014)

---

I always thought it was more like:

Those are the fundamental building blocks of relational databases. A primary key uniquely identifies a row within the database an individual table....

Where it says "relational databases" is a link. There is another at "primary key" and "foreign key" in the second paragraph.

You're hard coding the value as 4. Within a script as you've outlined, you'd have to change the value to 5, 6 or 7 to get those values. If you...

I also agree with Spaghetti on this one. Your approach will work. It's just not very clear. A lack of clarity can absolutely lead to error. For example, with only...

Those are the fundamental building blocks of relational databases. A primary key uniquely identifies a row within a table. A foreign key is used to relate one table to another,...

There is no specific to SQL Server function called a "refresh." Most of the time I've seen it used it's referring to a process (created by a business) wherein an...

That's really odd that the 120 compatibility level would do this. Have you applied any of the CUs?

My first thoughts were around parameter sniffing and issues there, especially if you're seeing changes in behavior based on changing the cardinality estimator (changing the compatibility level). But, I'm also...

Put your database code into source control and start treating it, as much as possible, like your application code. But, if you really want to start talking about database deployment...

Take a look at what Kendal Van Dyke has been working on, it might help.

You might take a look at this article[/url] on Simple-Talk. It could help.

That's pretty small data growth. Check the execution plans for the query from the old database as compared to the new one. See if they're different.

One very good reason for calling procedures from within a wrapper procedure is when you're dealing with IF statements in the procedure that would lead to bad parameter sniffing issues...

With completely different data sets? No, there's no way to use transaction logs to do this. What you're describing is basically merge replication. That might be something to explore.