On Sun, Mar 22, 2015 at 4:27 PM, Igor Tandetnik <igor at tandetnik.org> wrote:
> On 3/22/2015 11:50 AM, Bart Smissaert wrote: > >> Still, in this particular case it seems odd as there is only one column >> and >> one table in the sub-select. >> > > I'm not sure I understand what significance you ascribe to this fact. Why > again should the number of columns or tables in subselect matter? Thinking back to when I was first learning SQL, I remember being surprised in a similar way. To my procedural / object oriented / imperative way of thinking, the subselect was like a function, and I didn't expect the "namespace pollution" (which isn't really namespace pollution, but that's how it seemed at the time). -- Scott Robison
