The problem that the author had to fix had nothing to do with identity columns. This could have been a regular int column that got its new numbers via time-tested "max(colval) + 1" and it still would have run into the same issue. It's funny how identities have somehow gotten this bad rap. I've seen it everywhere I've gone in my career. In reality, identity columns are awesome when the DBA knows how to design them into the schema to match the business logic and expected record-load. They save a ton of index-scans (or table scans for some DBAs who use poorly designed indexes).

On the flip side, it is good the author featured a number of ways to list the identity columns in the database, which is helpful. The fix must have been easy, right? I mean, you can alter an int identity up to a bigint in one statement.

-Vic