On Mon, Feb 15, 2016 at 2:05 PM, Simon Slavin <slavins at bigfraud.org> wrote:
> > On 15 Feb 2016, at 8:58pm, James K. Lowden <jklowden at schemamania.org> > wrote: > > > Clemens Ladisch <clemens at ladisch.de> wrote: > > > >> I don't know why correlated subqueries cannot use values from the > >> outer query in the ORDER BY or LIMIT clauses; > > > > ORDER BY is not part of SELECT! It's not a relational operator. > > Per the SQL standard -- ORDER BY cannot appear in a subquery. It can > > appear in only one place: as a kind of post-processor that determines > > the order in which the rows are delivered to the client. > > And to build on that, if you cannot specify ORDER BY then you cannot > expect LIMIT to do anything useful. Since you don't know which ones will > be top in the list there's no point in picking just the top ones. > I think there is still a potential use for limit (though maybe there's another way I haven't considered). You have a table representing a job queue. Each job will take some amount of time to process. The order doesn't matter as long as all jobs are eventually processed and you have a single process running the jobs. Limit 1 is a reasonable way to grab a single job. -- Scott Robison
