OR is indeed not bad, but you have to think about when to use it. It can easily make queries slower compared to other operators. That’s exactly why we have the terms "ugly OR" / "bad OR".
It’s worth having a mental model of _why_ OR can be suboptimal. Often it’s because you’re only hitting an index on half the conditional, or forcing PG into a bitmap scan, or worse turning and index lookup into a sequential scan. Not to bang on about indexes too much but a partial index on the OR condition works if you’re lazy, although splitting into two queries is often a great solution as it gives you two fast queries to combine instead of one slow one (although sometimes that implies PG’s statistics are incorrect because it might have been able to know which side of the conditional cuts out more data).
