Question

Currently I am using MS Sql2000, though we are discussing upgrading this to 2005 (if that affects anything, Im guessing that it is pretty standard SQL that I need)

One of our products Tracks Sales from various departments around the country.
Currently, I run 3 Almost identical Stored Procedures in Query Analyzer, and then copy the results into excel, and spend an hour manipulating it all onto one spreadsheet.

What I’d like to do is to come up with a solution for combining all my stored procedures so that I end up with a single, combined Dataset I can just drop into excel and mail to the interested parties.

The solution I started to come up with involved a temp table, where I would first pull in all the sites, and then in turn update this with the results from each of the queries I currently run independently, but I got into all kinds of trouble, so I figured better to start again, and to ask for guidance and help from the off.

SQL is kind of paraphrased so please dont pick me up on simple syntax errors 🙂

A much simplified set of data would be;
table site {
id, location}

table salesEnquiry {siteId, custoemrId, 
DepositDate, BalancePaidDate, Status} 
--Status can be (0,1,2,3 for not started, deposit, completed,cancelled

And from that, I am thinking along the lines of;

--Set up the Whole Table
declare @tmp TABLE (id, Site, Deposits, Completed, Cancelled) 

insert into @tmp (id, site)
select id, site from site

update @tmp
set Deposits = (select count(customerId) from salesEnquiry
    where DepositPaidDate >= @startOfMonth and DepositPaidDate <= @endOfMonth and status = 1)  

** This is the point I get lost I think, as Im not sure how to relate total number of deposits taken against the tmpId's as not all departments will take a deposit each month.

I would then repeat the above methodology for Completed and Cancelled Customers so I have a final table like below.

SiteId    Location    TotalDeposits    TotalCompleted    TotalCancelled    
  1          a             3               0                  1
  2          b             0               0                  0
  3          c             1               17                 0

edited to add sample data + Outcome.
** Sorry, cant figure the formatting **

Ok, hope this helps.
site
id location
1 a
2 b
3 c

salesEnquiry
id   custId   deposit      complete     status
1    1        10/05/2011   null         1
2    2        11/04/2011   11/05/2011   2
1    3        12/05/2011   null         1
1    4        13/05/2011   13/05/2011   2
3    5        14/05/2011   null         3
3    6        13/02/2011   13/05/2011   3

would give

SiteId Location Deposits Completed Cancelled
1        a       2          1        0
2        b       0          1        0
3        c       0          0        2

Answer 1

I’m not sure what the “and so on” indicates, but up until that point the following should work:

SELECT
    S.id AS SiteId,
    S.site AS Location,
    SUM(CASE WHEN SE.status = 1 THEN 1 ELSE 0 END) AS TotalDeposits,
    SUM(CASE WHEN SE.status = 2 THEN 1 ELSE 0 END) AS TotalCompleted,
    SUM(CASE WHEN SE.status = 3 THEN 1 ELSE 0 END) AS TotalCanceled
FROM
    Sites S
LEFT OUTER JOIN SalesEnquiry SE ON
    SE.customer_id = S.id AND
    (
        (
            SE.status IN (1, 3) AND
            SE.DepositPaidDate >= @startOfMonth AND
            SE.DepositPaidDate <= @endOfMonth
        ) OR
        (
            SE.status IN (2, 3) AND
            SE.BalancePaidDate >= @startOfMonth AND
            SE.BalancePaidDate <= @endOfMonth
        )
    )
GROUP BY
    S.id,
    S.site
ORDER BY
    S.id

I had to guess a bit on the JOIN condition. You’re subquery in your question doesn’t seem to be relating the Site and SalesEnquiry tables. Was it really giving you the correct numbers? Join them on whichever column is relevant. I guessed the customer_id and id.