David,

I'd say that I owe you an apology!  Your code does, in fact, return all of the rows expected!  I don't know what I...

Ugh... No cursors... although it can certainly be done with one ...

BJure,

When you say you tested "IT" and NOT EXISTS was faster, what is "IT"?  In other words, what was "NOT EXISTS" faster than?  Thanks.

Hey Matt,

I get it... when sorted in the order of the first column, you want to show only those rows whose value in the third column has changed from the...

All of the methods about system tables and procedures that begin with "sp_" are great... but if all you want to do is get all the columnar info to show...

Cmore,

Here's one more little "secret" of mine ... I never use BCP to import to a "live" table

It, of course, depends on the SELECT statement in the NOT EXISTS but, I'd have to say "Yes", in most cases (was going to say ALWAYS but, sure-as-shootin', somebody would...

Just a note... The IDENTITY function can only be used when the INTO clause (as Figaro did) is also present or you will get the following error...

Server: Msg...

Haven't tested this but should work just fine.  Notice how DATEADD was used to subtract 4 months from the current date... not a job for DATEDIFF as many would think...

It won't actually report what constraint was violated but it will reject any record to the "error log" that either violates a constraint, key, foreign key, data-type, or column width. ...

Thought I'd clarify my statement about using an Outer Join as a replacement for a "Where Not In"...

About outer joins... I'm sure you've heard all sorts of horror...

If you are using an outer join as a replacement for a "WHERE NOT IN", then the filter must be in the WHERE clause as in Gary's second example above.

Use CASE as a substitute for DECODE.  Of course, "equivalent keywords" usually are not "equivalent" and will require a bit of rework.

Don't know what DISTINCTROW is... and don't want to...

Howdy,

It is and always has been a "bug" with Enterprise Manager.  I don't think you'll find a fix for it even on the Microsoft site (I've look for this fix...

Steve,

To answer your first question, I've found that PatIndex is a wee bit faster than LIKE.  In either case, both take a toll on speed when it comes to large data...