12

:) I have 2 queries, and I need to join them, I need to compare the working time of employee depending on activity with total working time of company in the same activity in defined period

First query is:

SELECT u.login,
       a.article, 
       p.p_article, 
       (SUM(p.p_going) + SUM(p.p_leaving) + SUM(p.p_working)) AS tottime
FROM pos p,users u, articles a
WHERE u.login = p.p_login
AND REPLACE( u.login, '.', '_' ) = 'users_name'
AND p.p_datum >= '2013-04-09'
AND p.p_datum <= '2013-04-16'
AND p.p_article = a.id
GROUP BY a.article

And my second query is:

SELECT a.article, 
       p.p_article, 
       (SUM(p.p_going) + SUM(p.p_leaving) + SUM(p.p_working)) AS tottime
FROM pos p, articles a
WHERE p.p_datum >= '2013-04-09'
AND p.p_datum <= '2013-04-16'
AND p.p_article = a.id
GROUP BY a.article

The first query returns me total working time of WORKER grouped by activites, for example:

u.login    a.article     p.p_article  tottime
Ivan       Teambuilding    1          3,45
Julie      Social_work     2          5,67

The second query returns me total working time of COMPANY grouped by activites, for example:

a.article     p.p_article  tottime
Teambuilding    1         150
Social_work     2         260

I want to have something like this, so I can compare the total time of worker per activity with total time of company working hours per activity in specific period:

u.login    a.article     p.p_article  tottime(worker)  tottime(company) 
Ivan       Teambuilding    1          3,45              150  
Julie      Social_work     2          5,67              260

In case of the NULL values I would like to use LEFT JOIN. I was searching for the solution for 3 hours, and everything I try is not working, so any help would be appreciated.

enigmaticus
  • 548
  • 3
  • 8
  • 26

2 Answers2

13

You can just join the 2 queries together as a pair of subselects.

Something like:-

SELECT Sub1.a, Sub1.b, Sub2.c
FROM (SELECT a, b FROM z) Sub1
INNER JOIN (SELECT a, c FROM y) Sub2
ON Sub1.a = Sub2.a

However can't really give you more as you first example query doesn't seem to bring back the details you say (only brings back 3 columns).

EDIT - With the corrected queries

SELECT Sub1.login AS User_name, Sub1.article AS Activity, Sub1.p_article AS `Activity id`, Sub1.tottime AS `Totaltime(worker)`, Sub2.tottime AS `Totaltime(company)`
FROM (SELECT u.login,a.article, p.p_article, (SUM(p.p_going) + SUM(p.p_leaving) + SUM(p.p_working)) AS tottime
FROM pos p
INNER JOIN users u ON u.login = p.p_login 
INNER JOIN articles a ON p.p_article = a.id
WHERE REPLACE( u.login, '.', '_' ) = 'users_name'
AND p.p_datum >= '2013-04-09'
AND p.p_datum <= '2013-04-16'
GROUP BY a.article) Sub1
INNER JOIN 
(SELECT a.article, p.p_article, (SUM(p.p_going) + SUM(p.p_leaving) + SUM(p.p_working)) AS tottime
FROM pos p
INNER JOIN articles a ON p.p_article = a.id
WHERE p.p_datum >= '2013-04-09'
AND p.p_datum <= '2013-04-16'
GROUP BY a.article) Sub2
ON Sub1.p_article = Sub2.p_article
Kickstart
  • 21,403
  • 2
  • 21
  • 33
4

The simplest would be using sub-queries (though they're in general not too efficient, but those GROUP BY's may make other solutions difficult).

Something like this should do it:

SELECT a.*, b.tottime AS 'Total time (company)'
FROM
    (SELECT u.login, a.article, p.p_article, (SUM(p.p_going) + SUM(p.p_leaving) + SUM(p.p_working)) AS 'Total time (worker)'
    FROM pos p, users u, articles a
    WHERE u.login = p.p_login
    AND REPLACE( u.login, '.', '_' ) = 'users_name'
    AND p.p_datum >= '2013-04-09'
    AND p.p_datum <= '2013-04-16'
    AND p.p_article = a.id
    GROUP BY a.article) a
LEFT JOIN
    (SELECT a.article, p.p_article, (SUM(p.p_going) + SUM(p.p_leaving) + SUM(p.p_working)) AS tottime
    FROM pos p, articles a
    WHERE p.p_datum >= '2013-04-09'
    AND p.p_datum <= '2013-04-16'
    AND p.p_article = a.id
    GROUP BY a.article) b
ON a.article = b.article /* AND a.p_article = b.p_article ?? */
Bernhard Barker
  • 54,589
  • 14
  • 104
  • 138
  • just edited :) thank you for help, I was trying all the time with something similar as you've done it here, but it didn't work :( so thanks a lot! :) – enigmaticus Apr 23 '13 at 08:20