Question

I have to generate a report using 4 tables, I tried a lot of joins, but I get wrong values 🙁

Hope someone will help me.

StaffAssign Table:

----------------------------
| Tid | StaffId | Planned  |
----------------------------
| 123 |   2     | 10:00:00 |
| 123 |   4     | 05:00:00 |
| 124 |   2     | 09:00:00 |
----------------------------

ActualEffort Table:

--------------------------------------------
| Tid | StaffId | ActualEffort  | logdate  |
--------------------------------------------
| 123 |   2     | 05:00:00      |2012-09-01|
| 123 |   4     | 05:00:00      |2012-09-01|
| 123 |   2     | 06:00:00      |2012-09-03|
| 124 |   2     | 09:00:00      |2012-09-04|
--------------------------------------------

ProjectList Table:

-------------------
| ProjectId | Tid |
-------------------
|    1      | 123 |
|    2      | 124 |
-------------------

ProjectName Table:

-------------------
| Id  | name      |
-------------------
| 1   | project1  |
| 2   | project2  |
-------------------

Below is the Query I’m using:

SELECT P.id AS projectid,P.name,
SEC_TO_TIME(SUM(TIME_TO_SEC(A.planned)))planned,SEC_TO_TIME(SUM(TIME_TO_SEC(E.actualeffort)))actual FROM actualeffort E 
INNER JOIN staffassign A ON A.tid = E.tid
INNER JOIN projectlist L ON L.tid = A.tid
INNER JOIN projectname P ON P.id = L.projectid 
WHERE E.logdate BETWEEN FROM_UNIXTIME('1346475600') AND FROM_UNIXTIME('1348290000')
GROUP BY P.id

I’m getting this:

---------------------------------------------
| projectid | name     | planned | actual   |
---------------------------------------------
|     1     | project1 | 15:00:00| 12:00:00 |
|     2     | project2 | 09:00:00| 09:00:00 |
---------------------------------------------

But it should be like:

---------------------------------------------
| projectid | name     | planned | actual   |
---------------------------------------------
|     1     | project1 | 15:00:00| 16:00:00 |
|     2     | project2 | 09:00:00| 09:00:00 |
---------------------------------------------

I’m very confused that, I don’t know where I’m going wrong with the joins.

Someone please help me, I’m struck with this.

Answer 1

The WHERE filter may be wrong. Note that first datetime in the condition starts at ’08:00:00′ –

SELECT FROM_UNIXTIME('1346475600') dt1, FROM_UNIXTIME('1348290000') dt2;
+---------------------+---------------------+
| dt1                 | dt2                 |
+---------------------+---------------------+
| 2012-09-01 08:00:00 | 2012-09-22 08:00:00 |
+---------------------+---------------------+

The type of ActualEffort.logdate field is DATE; the value ‘2012-09-01’ is less then ‘2012-09-01 08:00:00’. So change this condition or use DATE function –

WHERE
  E.logdate BETWEEN
  DATE(FROM_UNIXTIME('1346475600')) AND DATE(FROM_UNIXTIME('1348290000'))

---

(edited)

For one thing, when you join StaffAssign and ActualEffort on the Tid column only, you get a mini-Cartesian product for Tid=123 in particular, because both tables contain more than one row with that Tid and there’s no other specific condition to establish a 1:1 or, at most, 1:N relationship between the rows.

But in fact, 1:N wouldn’t do either: although it wouldn’t give you the Cartesian product effect, it would result in duplicated values on one side and, as a consequence, in a distorted SUM.

Therefore, data in StaffAssign and those in ActualEffort must be aggregated separately, then joined, something like this:

SELECT
  pl.ProjectId,
  pn.name,
  sa.Planned,
  ae.ActualEffort
FROM ProjectList pl
  JOIN ProjectName pn
    ON pn.Id = pl.ProjectId
  LEFT JOIN (SELECT Tid, SEC_TO_TIME(SUM(TIME_TO_SEC(Planned))) Planned FROM StaffAssign GROUP BY Tid) sa
    ON sa.Tid = pl.Tid
  LEFT JOIN (SELECT Tid, SEC_TO_TIME(SUM(TIME_TO_SEC(ActualEffort))) ActualEffort FROM ActualEffort GROUP BY Tid) ae
    ON ae.Tid = pl.Tid
 GROUP BY ProjectId;