Combine two queries - Postgresql - postgresql

Question

I am working on an sql query where I need the results from two queries. The first query is this:
select s.code, count(s.code), s.name from course_staff c join courses e on
c.course = e.id
join subjects s on e.subject = s.id
join semesters x on e.semester = x.id
where c.staff = 5033690
group by s.code, s.name
having count(s.code) > 8;
This results in the following:
code | count | name
----------+-------+--------------------------------
MATS1464 | 10 | Professional Comm&Presentation
MATS6605 | 9 | Prof Comm & Presentation
The second query:
select count(distinct s.code) from course_staff c join courses e
on c.course = e.id
join subjects s on e.subject = s.id
join semesters x on e.semester = x.id
where c.staff = 5033690
having count(s.code)>20;
This results in the following:
count
-------
30
I need to combine both the having count conditions from the above two queries into one query so that I can use it in a function or view. Please advise on how can I join these two queries so that both the having conditions are met. Help is much appreciated. Thanks

Answer 1

Here's an idea (using Postgres' Window Functions):
WITH
code_name (code, name) AS (
SELECT s.code, s.name
FROM course_staff c
JOIN courses e ON c.course = e.id
JOIN subjects s ON e.subject = s.id
JOIN semesters x ON e.semester = x.id
WHERE c.staff = 5033690
),
code_name_counts (code, name, count_both, count_total) AS (
SELECT DISTINCT code,
name,
COUNT(code) OVER (PARTITION BY code, name),
COUNT(code) OVER (PARTITION BY code)
FROM code_name
)
SELECT code, name, count_both, count_total
FROM code_name_counts
WHERE count_both > 8 AND count_total > 20 ;
Consider it a shot in the dark without further knowledge about your requirements.
And for testing purposes (without the WHERE):
WITH
code_name (code, name) AS (
VALUES (1, 'a'), (1, 'b'), (2, 'b'), (2, 'b'), (3, 'c')
),
code_name_counts (code, name, count_both, count_total) AS (
SELECT DISTINCT code,
name,
COUNT(code) OVER (PARTITION BY code, name),
COUNT(code) OVER (PARTITION BY code)
FROM code_name
)
SELECT code, name, count_both, count_total
FROM code_name_counts ;
which gives
code | name | count_both | count_total
------+------+------------+-------------
1 | a | 1 | 2
2 | b | 2 | 2
1 | b | 1 | 2
3 | c | 1 | 1
And here is an example without DISTINCT and PARTITION BY:
WITH
code_name (code, name) AS (
VALUES (1, 'a'), (1, 'b'), (2, 'b'), (2, 'b'), (3, 'c')
),
code_name_distinct (code, name) AS (
SELECT DISTINCT code, name
FROM code_name
),
code_name_counts (code, name, count_both, count_total) AS (
SELECT code,
name,
(SELECT COUNT(*) FROM code_name a
WHERE a.code = x.code AND a.name = x.name),
(SELECT COUNT(*) FROM code_name b
WHERE b.code = x.code)
FROM code_name_distinct x
)
SELECT code, name, count_both, count_total
FROM code_name_counts
which returns the same result as above.