I'm not really looking for information on how to speed this query
up. I'm just trying to interpret the output enough to tell me which step
is slow:
Seq Scan on mags (cost=0.00..187700750.56 rows=47476 width=4) (actual time=3004851.889..3004851.889 rows=0
loops=1)
Filter: ((signum IS NOT NULL) AND (NOT (SubPlan 1)))
SubPlan 1
-> Materialize (cost=0.00..3713.93 rows=95862 width=4) (actual time=0.011..16.145 rows=48139 loops=94951)
-> Seq Scan on sigs (cost=0.00..2906.62 rows=95862 width=4) (actual time=0.010..674.201 rows=95862
loops=1)
Total runtime: 3004852.005 ms
It looks like the inner seq scan takes 674ms, then the materialize takes
an additional 16ms? Or is that 16ms * 94951? Or 674 * 94951?
And the outer seq scan takes 3004851-3004851 = 0ms?
