如何将多个 SQL 选择合二为一

Question

So I have multiple tables that have a similar format to below:

user
----
user_id
first_name
last_name
group_id

group
-----
group_id
group_name

meeting
-------
meeting_id
group_id
user_id
meeting_date
completed
cancelled

I want to do a select to get all the meetings a user has completed by year, to far I have been doing 2 separate queries get all the users then loop through and get all their meetings on an annual basis:

select 
     u.id, u.first_name, u.last_name, g.name as groupname, g.id as group_id, 
from 
     users u, group g left join group g on u.group_id=g.group_id
order 
     by g.name asc, u.first_name asc

then a simple loop and foreach user_id I do the following:

select yearname, SUM(total_meetings) as total_meetings
        from  ( 
         select 
                 YEAR(meeting_date) as yearname, count(*) as total_meetings
         from
                 meetings
         where 
             user_id='value_from_foreach_loop'
             completed='1' and 
             YEAR(meeting_date) > 2016 
         group by 
             YEAR(time_of_event)
       ) t
group by 
     yearname   

I end up with data I can combine into a CSV ie:

u.id, u.first_name, u.last_name, group_name, group_id, yearname,  total_meetings
================================================================================================
1,    bob,          Roberts,     sales,      2,        2017,      248
1,    bob,          Roberts,     sales,      2,        2018,      988
1,    bob,          Roberts,     sales,      2,        2019,      848
2,    bill,         Henry,       sales,      2,        2017,      123
2,    bill,         Henry,       sales,      2,        2018,      1
2,    bill,         Henry,       sales,      2,        2019,      23

What I really need to do is combine these into one query, I need to remove the code that is executing the query and run it as straight SQL. I am stuck trying to determine how the sis possible.

I am sorry if this is really obvious I haven't actually touched SQL in many years and I just inherited a project and I am not sure where I start on this.

Thanks in advance.

Answer 1

Something like this:

select 
    u.id, u.first_name, u.last_name, g.name as groupname, g.id as group_id, YEAR(m.meeting_date), COUNT(*) as total_meetings
from 
    users u
    inner join `group` g on u.group_id = g.group_id
    inner join meetings m on u.id = m.user_id
where
    m.completed = 1 and 
    m.meeting_date > '2016-01-01'
group by
    u.id, u.first_name, u.last_name, g.name, g.id, YEAR(m.meeting_date)

You don't need your double grouping counting/summing of years, and anything that you select and then run in a loop where you're just plugging an id into a where clause is a candidate for doing in the DB with a join

Answer 2

Implicit joins, as @CaiusJard pointed out, have been deprecated for roughly the last 30 years. Implicit joins were formulated as a comma separated list of tables/views, and the join conditions went into the WHERE condition.

We nowadays use explicit joins throughout. Even cross joins should explicitly be expressed with the CROSS JOIN keywords. And don't use "group" as an object name if you can avoid it. You would run into a syntax error almost in any database, as it is one of the most important keywords in SQL. I put it into double quotes in my example.

Having said that - all you have to do is to join , then group in one SELECT :

SELECT
  u.id
, u.first_name
, u.last_name
, g.name AS groupname
, g.id AS group_id
, YEAR(m.meeting_date) AS yearname
, COUNT(*) AS total_meetings
FROM
  users u
JOIN meetings m ON u.id=m.user_id AND m.completed=1
LEFT JOIN "group" g ON u.group_id = g.group_id
GROUP BY
  u.id
, u.first_name
, u.last_name
, g.name 
, g.id 
, YEAR(m.meeting_date) 
ORDER BY
  g.name ASC
, u.first_name ASC 
;