多连接查询中的SQL总和

Question

I'm struggling for the setup for this. This is the setup of the first table. The first query needs to only apply to items past a certain date, say 2012-06-01.

|                  Table: TIMELOG                  |
| INQUIRY_ID | … |      LOGMINS    | … |  OPERID   | … |    EDATE    |
|     2      |   |       45        |   |  bob      |   | 2012-05-01  |
|     5      |   |       3         |   |  richard  |   | 2012-06-02  |
|     5      |   |       12        |   |  bob      |   | 2012-07-01  |
|     5      |   |       15        |   |  paul     |   | 2012-07-01  |
|     6      |   |       10        |   |  paul     |   | 2012-07-01  |

The query then needs to go into a different table using INQUIRY_ID and get some additional fields (there is only one entry per INQUIRY_ID in this next table:

|                  Table: INQUIRY                  |
| INQUIRY_ID | … |  CATEGORY_ID    | … |  PROD_ID  |
|     2      |   |       45        |   |  6        |
|     3      |   |       3         |   |  50       |
|     4      |   |       12        |   |  3        |
|     5      |   |       15        |   |  67       |
|     6      |   |       10        |   |  2        |

Once we have the CATEGORY_ID and PROD_ID the "friendly names" for those are in their own seperate tables

|            Table: CATEGORY            |
| CATEGORY_ID | … |      CATNAME        |
|     45      |   |       Server        |
|     3       |   |       Workstation   |
|     12      |   |       Phones        |
|     15      |   |       Backup        |  
|     10      |   |       Network       |

|            Table: PROD                |
| PROD_ID     | … |      PRODDESC       |
|     6       |   |       SBS 2003      |
|     50      |   |       Windows 7     |
|     3       |   |       iPhone        |
|     67      |   |       Buexec        |  
|     2       |   |       SwitchF       |   

I'd then like to total the LOGMINS field, bearing in mind sometimes it will have multiple LOGMINS per INQUIRY_ID To finish up as

| INQUIRY_ID | … |      TOTAL      | … |  CATNAME         | … |   PRODDESC    |
|     2      |   |       45        |   |     Server       |   |   SBS 2003    |
|     5      |   |       30        |   |     Workstation  |   |   Windows 7   |
|     6      |   |       10        |   |     Phones       |   |   iPhone      |

I'm sorry for the extraordinary amount of detail, the queries I have tried just bomb out, my skill in SQL blows up when I get into nested queries and multiple joins. Any help would be massively appreciated.

Query I am running now:

 select 
 timelog.INQUIRY_ID,
 SUM (logmins) AS Total,
 catname,
 proddesc,
 EDATE
from
 timelog
inner join inquiry on timelog.inquiry_id=inquiry.inquiry_id
inner join category on inquiry.category_id = category.category_id
inner join prod on inquiry.prod_id = prod.prod_id
where EDATE > '2013-07-01'
group by
timelog.INQUIRY_ID,
catname,
proddesc,
edate
order by timelog.INQUIRY_ID desc

Answer 1

Join the tables your data is in together, then group on the data that you are keeping constant, and sum the fields you want to add

select 
  inquiry.inquiry_id,
  SUM(logmins),
  catname,
  proddesc
from
   timelog
    inner join inquiry on timelog.inquiry_id=inquiry.inquiry_id
    inner join category on inquiry.category_id = category.category_id
    inner join prod on inquiry.prod_id = prod.prod_id
group by
  inquiry.inquiry_id,
  catname,
  proddesc