insert missing data into a result (mysql)

Question

I have the following table :

id	site	date	id_action	id_customer	id_sim	id_plq
7	mmg	2020-10-31 15:34:19	9	119
8	mmg	2021-10-31 15:34:29	9	119
11	mmg	2021-10-31 15:34:29	10	118
12	chrono	2021-10-31 18:31:29	9	118
13	mmg	2021-10-31 20:08:25	9	118
14	mmg	2021-09-30 21:09:14	9	118
15	plx	2020-10-30 21:09:14	9	119
16	plx	2021-10-31 22:12:45	12	118
17	plx	2020-10-31 22:12:45	9	118
18	plx	2021-10-31 22:12:45	9	118
19	plx	2021-08-31 22:12:45	9	118
20	plx	2021-10-31 22:12:45	9	119
21	plx	2021-07-27 22:12:45	9	116
22	plx	2021-10-31 22:12:45	9	118
23	plx	2021-06-28 22:12:45	9	118
24	plx	2020-10-31 22:12:45	9	119
25	plx	2020-06-30 22:12:45	9	118
26	plx	2020-10-31 22:12:45	9	118
27	plx	2020-10-31 22:12:45	9	119
28	plx	2020-10-31 22:12:45	9	118
29	plx	2020-08-31 22:12:45	9	118
31	chrono	2021-10-31 18:31:29	9	116

Actually I do the following request :

select YEAR(b.date) as year,
       MONTH(b.date) as month,
       b.site,
       count(distinct b.id) as 'totalLogin/Month',
       count(distinct b.ID_customer) as 'totalUser/Month'
from awm_logs b
group by year, month, site
order by year desc, month desc, site;

This give me this result :

year	month	site	'totalLogin/Month'	'totalUser/Month'
2021	10	chrono	2	2
2021	10	mmg	3	2
2021	10	plx	4	2
2021	9	mmg	1	1
2021	8	plx	1	1
2021	7	plx	1	1
2021	6	plx	1	1
2020	10	mmg	1	1
2020	10	plx	6	2
2020	8	plx	1	1
2020	6	plx	1	1

It is possible to insert the missing month in the result ? Without another table ?

And possibly the missing site ?

My idea of the result :

for avoiding long table i made the example with just the two first month

year	month	site	'totalLogin/Month'	'totalUser/Month'	note
2021	10	chrono	2	2
2021	10	mmg	3	2
2021	10	plx	4	2
2021	9	chrono	0	0	new line
2021	9	mmg	1	1
2021	9	plx	0	0	new line

Answer 1

Build another table of all the needed dates, then do

SELECT ...
FROM dates AS d
LEFT JOIN your_table AS t ON t.date = d.date
....

dates could be manually created once, then kept for use. Or write code to generate. Or switch to MariaDB with its seq_1_to_100 type pseudo table.