just for curiosity - why the HAVING clause was not used?
Any window functions are +/- an "aggregate" function, and then HAVING looks more natural to me.
Hm, HAVING requires to apply 'group by' which windows functions do not require (unlike aggregates).
superuser@postgres=# select * from (select 1 as v) q having true limit 1; ERROR: column "q.v" must appear in the GROUP BY clause or be used in an aggregate function LINE 1: select * from (select 1 as v) q having true limit 1;
If a query has both window function and grouped aggregate, HAVING would be applying at different grains potentially? If so, seems sus.
