roblew 15918 (3/21/2014)

Thanks,...

Backing up Gail, piling on and reinforcing this one.

REPAIR_ALLOW_DATA_LOSS is extremely dangerous. It will attempt to fix the database without any regard to your data at all. You could lose...

If you don't mind spending a little money, Red Gate SQL Prompt[/url] is a tool that I can't live without. It formats your code. And it provides code completion way...

If, logically, you need all the rows from one table, but only the matching rows from another table, then some type of OUTER JOIN is necessary. They're not inherently evil.

VMs are pretty solid and work quite well. Just as well as physical machines these days if they're properly set up and maintained. But, performance could be an issue there...

Other than what Gail outlines, which is a good way to get a new plan every time, why are you trying to "ignore the cache"?

Can you tell what is slowing down the inserts? I mean other than the obvious, this index sucks. What does the execution plan look like? What are the wait stats...

I'm curious what you mean by administration commands? Except for the direct manipulation of data within the tables (and even that) almost every single command is going to be an...

If it's integer data, use the integer data type. Done.

Numbers in general are stored basically the same way within SQL Server. The only differences really are in the size and...

Rebuilding the log file requires that no open transactions exist in the database. Try deleting the log file and attaching the database without the log. It might not work, but...

Take a look at the estimated execution plan. It will tell you the choices SQL Server is making based on the code you provided and the statistics available on the...

Yeah, it also checks the file locations.

You need to do this:

RESTORE DATABASE MyNewDatabase

FROM DISK = 'x:\mybackupfile.bak'

WITH MOVE 'MyLogicalFileName' = 'd:\data\MyNewDatabase.mdf',

MOVE 'MyLogicalLogFileName' = 'e:\logs\MyNewDatabase.ldf';

You need to put in your appropriate paths...

It has to be either duplicate data or NULL values. With a unique index you can have a single NULL value. As was stated, with a primary key, you can't...

Rebuilding the index defragments the index, yes.

There's a bunch of stuff that's been on the deprecation list that's still not quite gone. Profiler and Trace Events are the big ones I'm curious about.