On Oct 12, 2015, at 10:39 AM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
>
> Jeff Janes <jeff.janes@gmail.com> writes:
>> On Mon, Oct 12, 2015 at 10:58 AM, Israel Brewster <israel@ravnalaska.net>
>> wrote:
>>> My first thought was to do something like this:
>>>
>>> SELECT * FROM (SELECT lognum,array_agg(flightnum) as flightnums FROM logs
>>> GROUP BY lognum) s1 WHERE '8%' like ANY(flightnums);
>>>
>>> But while this doesn't give an error, it also doesn't return any results.
>>> I'm guessing that this is because the wildcard is on the left of the
>>> operator, and needs to be on the right.
>
>> Right. The LIKE operator does not have a commutator by default. (And if
>> you created one for it, it could not use an index in this case.)
>
> Well, it couldn't use an index anyway, given that the query as written
> wants to collect groups if *any* member is LIKE '8%', rather than
> restricting the data to such flightnums before aggregation occurs.
>
> Personally I'd suggest building a commutator operator (just need a
> one-liner SQL or plpgsql function as infrastructure) and away you go.
That could work. I'll look into that.
>
>> I think you're best bet is to do a subquery against the unaggregated table.
>
>> select * from aggregated a where exists
>> (select 1 from unaggregated ua where a.lognum=ua.lognum and flightnum
>> like '8%')
>
> That would work too, but not sure about performance relative to the other
> way.
>
> regards, tom lane
-----------------------------------------------
Israel Brewster
Systems Analyst II
Ravn Alaska
5245 Airport Industrial Rd
Fairbanks, AK 99709
(907) 450-7293
-----------------------------------------------