So, Dave and I were having a conversation about clustered indexes over lunch yesterday.
Yeah, we're this boring.
Anyway, I have an issue at work where I have a table that has a horrible clustered index. It's not very selective (the first key can return upwards of 80K rows) and the first key changes frequently when the row is updated (leading to fragmentation). It's great for one semi-frequent query in our system, but really isn't useful for most others. The problem is that SQL Server (talking SQL Server 2000 here) uses the index A LOT in the form of clustered index scans. ...