I love joined subqueries in MySQL

So I was looking into ways of improving the load times of a project, and knowing SQL is one of my biggest weaknesses (in my own humble opinion), I figured there might be a fair bit to gain by researching a bit around it, which lead to a friend linking me to http://www.chrislondon.co/joins-vs-subqueries/

Judging by his benchmarks on a low amount of data points, he discovered a massive leap between a subquery and a joined subquery. I’ve been using join statements all along (and in some cases multiple queries like the schmuck I am) so I figured it was worth giving this a shot on the data I was working with. I mention the data points as it was put up in the comments that the amount of data he queried against was very low, I hit mine against ~50 000 rows of data

I made a dump of the live data from the system in production and started running queries.
The numbers I got out of it were quite amazing (I’ll leave the query I ran at the bottom for public scrutiny and humiliation).
The default view is to return 30 entries per page, with my old approach this took 0.6 seconds, which is a frighteningly long time in my opinion, even if it isn’t a front end system.
Then came the query I have below, the same 30 rows were returned in 0.02 seconds, a very substantial leap, so I got experimental, and I kept increasing the rows I pulled with the new query, wanting to see just how big of a leap I could make.
The final number? 5 000 rows of data to just barely break 0.6 seconds! That’s quite the improvement on the meager 30 rows it managed before, that’s for sure.
Now as you’ll notice, I only took the heavy parts of the query and put into subqueries for now, there are still a few regular join statements left in the query, I might experiment some more but didn’t want to mess up the entire system in one go.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
SELECT
	cd.triangel_course_id AS course_id,
	cd.triangel_course_courseid,
	cd.triangel_course_coordinator,
	cd.triangel_course_coordinator_2,
	cd.triangel_course_seats AS course_seats,
	cd.triangel_course_time_start AS course_start,
	cd.triangel_course_time_end AS course_end,
	cd.triangel_course_note AS course_note,
	cd.triangel_course_room,
	cd.triangel_course_price AS course_price,
	cd.triangel_course_price_private AS course_price_private,
	cd.triangel_course_email_note AS course_email_note,
	cd.triangel_course_english AS course_english,
	cd.triangel_course_web AS course_web,
	cd.triangel_course_complete_user AS course_complete,
	first_coordinator.triangel_instructor_name AS primary_coordinator,
	second_coordinator.triangel_instructor_name AS secondary_coordinator,
	c.courseid AS course_code,
	c.name AS course_name,
	c.name_en AS course_name_en,
	room.triangel_room_name AS classroom,
	tc.attendee_count,
	tc.invoiced_count,
	tr.reserved_count
FROM
	triangel_course cd
LEFT JOIN
	triangel_instructor first_coordinator
		ON ( first_coordinator.triangel_instructor_id = cd.triangel_course_coordinator )
LEFT JOIN
	triangel_instructor second_coordinator
		ON ( second_coordinator.triangel_instructor_id = cd.triangel_course_coordinator_2 )
LEFT JOIN
	courses c
		ON ( c.id = cd.triangel_course_courseid )
LEFT JOIN
	triangel_rooms AS room
		ON ( room.triangel_room_id = cd.triangel_course_room )
LEFT JOIN (
		SELECT
			COUNT(DISTINCT triangel_courses_id) AS attendee_count,
			COUNT(DISTINCT triangel_courses_faktura) AS invoiced_count,
			triangel_courses_course
		FROM
			triangel_courses
		WHERE
			triangel_courses_delisted <= 0
		AND
			triangel_courses_noshow = 0
		GROUP BY
			triangel_courses_course
	) tc
	ON tc.triangel_courses_course = cd.triangel_course_id
LEFT JOIN (
		SELECT
			SUM(triangel_courses_reserved_seats) AS reserved_count,
			triangel_courses_reserved_course
		FROM
			triangel_courses_reserved
		GROUP BY
			triangel_courses_reserved_course
	) tr
	ON tr.triangel_courses_reserved_course = cd.triangel_course_id
WHERE
	1 = 1

And yes, it’s terrible, SQL is far from my strong suit, and I’ll gladly be told better ways of approaching things in the query so do let me know!

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.