Query takes 20+ seconds, table is less than 300 rows - sql

Question

Here's the query:
with contrib as (
select
first_name,
last_name,
user_id,
photo_url
from contributors
where visible = true
group by 1,2,3,4
),
dwm as (
select * from dialogues_with_metadata
),
joined as (
select
c.*,
dwm.dialogue_id
from contrib c
left join dwm on c.user_id = dwm.contributor_one_user_id or c.user_id = dwm.contributor_two_user_id
)
select
first_name,
last_name,
user_id,
photo_url,
count(distinct dialogue_id) as dialogues
from joined
group by 1,2,3,4
order by 3 desc
PostgreSQL database.
CPU usage is 10%, so I don't think that's the problem!
I suspect what's slowing things down is the join statement. How might I reconfigure this query so that it doesn't take ~20 seconds to return less than 300 rows?
Here's the contributors table schema:
user_id (primary key - uuid)
username
first_name
last_name
hash
description
blurb
photo_url
blurb_updated_at
visible
And the dialogues table schema:
dialogue_id (uuid)
contributor_one_uuid
contributor_two_uuid
title
image_url
visible
categories
image_source
current_popularity
created_at
override_url
visible
Ran explain select * from dialogues_with_metadata; Here are the results:
Hash Left Join (cost=167.80..172.34 rows=137 width=880)
Hash Cond: (a.dialogue_id = b.writing_dialogue_id)
CTE main
-> Hash Right Join (cost=64.30..111.60 rows=137 width=504)
Hash Cond: (c2.user_id = d.contributor_two_uuid)
-> Seq Scan on contributors c2 (cost=0.00..43.60 rows=260 width=125)
-> Hash (cost=62.59..62.59 rows=137 width=325)
-> Hash Left Join (cost=46.85..62.59 rows=137 width=325)
Hash Cond: (d.contributor_one_uuid = c1.user_id)
-> Seq Scan on dialogues d (cost=0.00..15.37 rows=137 width=216)
-> Hash (cost=43.60..43.60 rows=260 width=125)
-> Seq Scan on contributors c1 (cost=0.00..43.60 rows=260 width=125)
CTE dialogues_with_installment_counts
-> HashAggregate (cost=50.39..52.00 rows=129 width=28)
Group Key: writings.dialogue_id
-> Seq Scan on writings (cost=0.00..46.82 rows=476 width=28)
Filter: finalized
-> CTE Scan on main a (cost=0.00..2.74 rows=137 width=868)
-> Hash (cost=2.58..2.58 rows=129 width=28)
-> CTE Scan on dialogues_with_installment_counts b (cost=0.00..2.58 rows=129 width=28)
EXPLAIN on updated query:
Seq Scan on contributors c (cost=0.00..4012.89 rows=247 width=109)
Filter: visible
SubPlan 1
-> Aggregate (cost=16.06..16.07 rows=1 width=8)
-> Seq Scan on dialogues (cost=0.00..16.05 rows=2 width=0)
Filter: ((c.user_id = contributor_one_uuid) OR (c.user_id = contributor_two_uuid))
explain (analyze, buffers) select on updated query:
Seq Scan on contributors c (cost=0.00..4205.86 rows=259 width=109) (actual time=0.073..16819.258 rows=260 loops=1)
Filter: visible
Rows Removed by Filter: 13
Buffers: shared hit=3681
SubPlan 1
-> Aggregate (cost=16.06..16.07 rows=1 width=8) (actual time=64.548..64.549 rows=1 loops=260)
Buffers: shared hit=3640
-> Seq Scan on dialogues (cost=0.00..16.05 rows=2 width=0) (actual time=49.155..64.547 rows=1 loops=260)
Filter: ((c.user_id = contributor_one_uuid) OR (c.user_id = contributor_two_uuid))
Rows Removed by Filter: 136
Buffers: shared hit=3640
Planning Time: 0.136 ms
Execution Time: 16819.365 ms
A second time (16 seconds faster!):
Seq Scan on contributors c (cost=0.00..4205.86 rows=259 width=109) (actual time=0.063..801.278 rows=260 loops=1)
Filter: visible
Rows Removed by Filter: 13
Buffers: shared hit=3681
SubPlan 1
-> Aggregate (cost=16.06..16.07 rows=1 width=8) (actual time=3.080..3.080 rows=1 loops=260)
Buffers: shared hit=3640
-> Seq Scan on dialogues (cost=0.00..16.05 rows=2 width=0) (actual time=0.009..3.079 rows=1 loops=260)
Filter: ((c.user_id = contributor_one_uuid) OR (c.user_id = contributor_two_uuid))
Rows Removed by Filter: 136
Buffers: shared hit=3640
Planning Time: 0.127 ms
Execution Time: 801.379 ms

Answer 1

This query should do the same without unnesesary grouping statements
select
c.first_name,
c.last_name,
c.user_id,
c.photo_url,
(
select
count(distinct dialogue_id) from dialogues_with_metadata dwm
where
c.user_id = dwm.contributor_one_user_id
or c.user_id = dwm.contributor_two_user_id
) as dialogues
from contributors c
where c.visible = true
Additionaly it's worth checking if counting number of conversations can be done in more efficient way (what indexes are on this table ?).
Query version with dialogues table
select
c.first_name,
c.last_name,
c.user_id,
c.photo_url,
(
select
count(*) from dialogues
where
c.user_id = dialogues.contributor_one_uuid
or c.user_id = dialogues.contributor_two_uuid
) as dialogues
from contributors c
where c.visible = true
Another version
select
c.first_name,
c.last_name,
c.user_id,
c.photo_url,
s.dialogues
from contributors c
join (
select count(*) dialogues, user_id from (
select contributor_one_uuid user_id from dialogues
union all
select contributor_two_uuid from dialogues
) stats group by user_id
) s on (s.user_id = c.user_id)
where c.visible = true