Question

Is there a way of populating the Cust1, Cust2, Cust3 columns in Table2 that doesn’t involve running the same query with several different “WHERE” clauses?

For example, right now I would resort to doing this:

INSERT INTO Table2(Cust1_Totals)
SELECT Fruit, SUM(quantity)
FROM Table1
WHERE Customer_ID LIKE 1
GROUP BY Fruit

INSERT INTO Table2(Cust2_Totals)
SELECT Fruit, SUM(quantity)
FROM Table1
WHERE Customer_ID LIKE 2
GROUP BY Fruit

INSERT INTO Table2(Cust3_Totals)
SELECT Fruit, SUM(quantity)
FROM Table1
WHERE Customer_ID LIKE 3
GROUP BY Fruit

Ideally, I would like to have another table with a Customer_ID column of all the unique customers and then I would reference each one in a loop in the above query. SQL Server 2008.

Answer 1

You could use case to generate sums per customer in separate columns:

INSERT  Table2
        (Produce, Cust1_Totals, Cust2_Totals, Cust3_Totals)
SELECT  Fruit
,       SUM(case when Customer_ID = 1 then quantity end)
,       SUM(case when Customer_ID = 2 then quantity end)
,       SUM(case when Customer_ID = 3 then quantity end)
FROM    Table1
WHERE   Customer_ID IN (1,2,3)
GROUP BY 
        Fruit

Generating a variable numbers of columns would require dynamic sql. It’s usually easier to let Excel that. In Excel-speak, it’s known as “pivoting”.