Getting product quantities broken down annually

Question

I'm fairly new to SQL and I'm trying to get my head around the following problem - I'm attempting to get a list of products broken down so we can see how many of each one has been sold every year

I have two tables; Orders (which contains JobNum and Date) and Products (with JobNum, Fig, Fype, Size and Quantity fields). Due to the products being custom made a product is an amalgamation of the fig, type and size fields. For example: To get a listing of all the product variations created I can do:

SELECT DISTINCT Fig,
                Type, 
                Size 
FROM Product;

And to get a list of quantities for all products made in a calendar year this query seems to work: (Date is the format 'dd-mm-yyyy' hence the LIKE '%2002' to filter by year)

SELECT product.fig, 
       product.type, 
       product.size, 
       Sum(product.qty) AS Quantity 
FROM   orders.dbf 
       INNER JOIN product.dbf 
               ON orders.jobnum = product.jobnum 
WHERE  orders.date LIKE '%2002' 
GROUP  BY product.fig, 
          product.type, 
          product.size 

Which gives the data for 1 year outputted like this:

Fig   Type   Size   Qty
 AA    B      2      1

My question is how can I pull out the data to get an output like this?

Fig   Type   Size  2001  2002  2003...
 AA    B      2     1     2     4
 BB    C      4     4     6     7

I can think how produce this in a program but I'd like to see if it's possible to do just in SQL?

Many thanks in advance.

EDIT- Can I just point out that Product.Type can be blank in some cases and in some years it is possible for zero instances of a product to be sold, so for that year the corresponding quantity amount could be blank or 0.

Answer 1

Although I have not specifically used dbase since about 1986, I am guessing they have implemented a common IIF() functionality as other languages. I am more used to FoxPro/Visual Foxpro historically, but here is the basis of what you need, and if you are using Visual Foxpro OleDB, you should be fine.

What is happening here, as you are going through, the group by will be based on each fig, type, size, no problem. Now, to create your PIVOT, I am just doing a sum( IIF()) based on the order date. If the year of the order date in question equals the respective year, add that quantity to that column's total, otherwise, sum a zero value.

SELECT
      p.fig,
      p.type,
      p.size,
      SUM( IIF( YEAR( o.date ) = 2002, p.qty, 0 )) as P2002Qty,
      SUM( IIF( YEAR( o.date ) = 2003, p.qty, 0 )) as P2003Qty,
      SUM( IIF( YEAR( o.date ) = 2004, p.qty, 0 )) as P2004Qty,
      SUM( IIF( YEAR( o.date ) = 2005, p.qty, 0 )) as P2005Qty,
      SUM( IIF( YEAR( o.date ) = 2006, p.qty, 0 )) as P2006Qty,
      SUM( IIF( YEAR( o.date ) = 2007, p.qty, 0 )) as P2007Qty,
      SUM( IIF( YEAR( o.date ) = 2008, p.qty, 0 )) as P2008Qty,
      SUM( IIF( YEAR( o.date ) = 2009, p.qty, 0 )) as P2009Qty,
      SUM( IIF( YEAR( o.date ) = 2010, p.qty, 0 )) as P2010Qty,
      SUM( IIF( YEAR( o.date ) = 2011, p.qty, 0 )) as P2011Qty,
      SUM( IIF( YEAR( o.date ) = 2012, p.qty, 0 )) as P2012Qty,
      SUM( IIF( YEAR( o.date ) = 2013, p.qty, 0 )) as P2013Qty
   from
      orders o
         INNER JOIN product p
            ON o.jobnum = p.jobnum
   group by
      p.fix,
      p.type,
      p.size

Now, one note... If the results for whatever strange reason gives zeros in most columns, it is because the engine is predetermining the column size width based on the first "test" record to run the query before it actually executes. If so, then two options to adjust... Change each row to one of the following

FROM:
SUM( IIF( YEAR( o.date ) = 2002, p.qty, 0 )) as P2002Qty,

TO: ( so it FORCES allocated space to 7 digit positions )
SUM( IIF( YEAR( o.date ) = 2002, p.qty, 0000000 )) as P2002Qty,

OR: (so it always uses the qty size, but multiplies by 1 or 0 so it still uses basis of qty column)
SUM( p.qty * IIF( YEAR( o.date ) = 2002, 1, 0 )) as P2002Qty,

Answer 2

You need to use PIVOT for transposing the rows to columns.

Microsoft link to read about PIVOT: http://technet.microsoft.com/en-us/library/ms177410(v=sql.105).aspx

For the year problem, there is an SQL function called YEAR(Date) which will pull out only the year from a given date. Eg:

select year('1-Jan-2013')

Answer 3

You may use a subquery as a column. For instance:

SELECT t0.*, (SELECT COUNT(*) FROM t1 WHERE YEAR(date_fld)=2002) AS y2002, (SELECT COUNT(*) FROM t1 WHERE YEAR(date_fld)=2003) AS y2003 FROM t0