Question

I am trying to have the results from two columns in table1 subtracted from one another and add the total from table2 should the data exist in that table. Not always will there be data in table2 so if there is nothing there I need to use “0”. Here is what I have so far which returns the wrong amount when there IS data in table2.

SELECT 
    CONVERT(CHAR(10),table1.PostingDate, 120) AS business_date, 
    table1.Location AS store_number, 
    (CASE WHEN COUNT(table2.Document) > 0 THEN 
          SUM(table1.Total - table1.TipAmount + table2.Total)
          Else 
                SUM(table1.Total-table1.TipAmount) 
    END) AS net_sales_ttl
FROM table1 
    left join table2 
        on CONVERT(CHAR(10),table1.PostingDate, 120) = CONVERT(CHAR(10),table2.PostingDate, 120)
WHERE table1.PostingDate between '2012-09-09' and '2012-09-16'
GROUP BY CONVERT(CHAR(10),table1.PostingDate, 120), table1.Location

Here is the results:

business_date   store_number    net_sales_ttl
2012-09-09  xxx         1699.61
2012-09-10  xxx             923.56
2012-09-11  xxx             1230.93  <--This should be 1399.93
2012-09-12  xxx             874.98
2012-09-13  xxx             1342.21
2012-09-14  xxx             1609.6
2012-09-15  xxx             2324.31

For some reason the query is not doing the math correctly and returning the wrong values. The only day that table2 has a value is 09-11-12 and that amount is only -1.00. It is giving me 1230.93 which is -169 from the correct value. I don’t know where the -169 is coming from when it should be -1.00. Original amount is 1400.93 in table1 and table2 should be subtracted from that which is -1.00 giving a result of 1399.93.

    Sample data:
    table1 has date, location, and sales
    09-09 1111 5.00
    09-10 1111 3.00
    09-11 1111 7.00
    09-12 1111 10.00
    table2 has refunds
    09-11 1111 -1.00

    Return set would look like this:
    09-09 1111 5.00
    09-10 1111 3.00
    09-11 1111 6.00   <--Reflecting the refund from table2
    09-12 1111 10.00

Answer 1

try

SELECT  
    CONVERT(CHAR(10),table1.PostingDate, 120) AS business_date,  
    table1.Location AS store_number,  
    SUM (table1.Total - table1.TipAmount + isnull(table2.Total,0))
FROM table1   
    left join table2  
    on CONVERT(CHAR(10),table1.PostingDate, 120) = CONVERT(CHAR(10),table2.PostingDate, 120)  
WHERE table1.PostingDate between '2012-09-09' and '2012-09-16'  
GROUP BY CONVERT(CHAR(10),table1.PostingDate, 120), table1.Location 

Also, some sample data would help – the totals alone can’t help

Based on your sample data, try this.

Select date, location, sum(sales)
from 
(
    select date, location, sales from table1 
    union
    select date, location, refunds from table2 
) v
group by date, location