I don't. I put distinct on everything by default, because it makes SQL more like relational algebra.
What's the argument for avoiding distinct? Is it performance? I've often wondered if the (sometimes redundant) distincts slow things down. Perhaps it does, but just not enough that I noticed.
What's the argument for avoiding distinct? Is it performance? I've often wondered if the (sometimes redundant) distincts slow things down. Perhaps it does, but just not enough that I noticed.