I apologise for the late response.
I've increased "effective_cache_size" to 50% and tried again - no change.
Afterwards, I've increased "cpu_tuple_cost" from 0.02 to 0.05 and tried again -
no change.
What is most curious to me is that I think the initial result set is very
small, so any JOIN
Sorry for the delay, but I didn't have access to the database during the
weekend.
Here's the output of "explain (analyze, buffers)":
http://explain.depesz.com/s/scC
I'm also curious why it actually seems to touch the table assuming there are
output columns which I haven't defined anywhere (s
Thanks for the suggestion.
I've tried it with seqscan set to off, but there's still a bitmap heap scan
going on:
http://explain.depesz.com/s/zIJl
I have random_page_cost set to 1.5 at the moment, as the database is on a solid
state disk.
Every user has a parent, but not every parent has a ch
Good day,
I have a recursive CTE where a table scan occurs, even though there doesn't
seem to be a good reason for it.
It seems the planner came to the conclusion that columns that are not actually
used in the output, joins or a where clause are a part of the output.
It's not a performance pro
