Get the two rows of customer with same amount value within consecutive year

Question

Hi,
Below are sample data, what I would like to do is pull the customer along with their two finance details that have the same amount within the two consecutive year. Please note that there are many customers and not all customers are have these criteria, so the query should not include those customers.
Like in my sample data the result should be:
(first row)
165321 John King Magazine 01/03/2007 30/03/2011 10/02/2009 180.00
(second row)
165321 John King Magazine     01/03/2007 30/03/2011 06/12/2008     180.00
Customer Table:
ID     Name Product     Start Delivery     End Delivery
165321     John King     Magazine     01/03/2007     30/03/2011
165322     Bryan Lee     Magazine     15/04/2002     30/04/2011
Finance Table:
ID          Invoice Date     USD Amount     
165321          10/02/2009     180.00
165321          10/02/2009     120.00
165321          06/12/2008     180.00
165321          20/05/2007     99.00
165321          12/03/2007     150.00
165321          07/03/2007     180.00
165322          01/01/2003     124.00
165322          30/05/2002     160.00
165322          20/04/2002     99.00
Thank you..

Answer 1

Hi,
Depending on your exact requirements, you can do something like this:
WITH     got_this_yr     AS
     SELECT    id
     ,       MIN (invoice_date)               AS invoice_date
     ,       usd_amount
     ,       EXTRACT (YEAR FROM invoice_date)     AS this_yr
     FROM       finance
--     WHERE       ...     -- any filtering goes here
     GROUP BY  id
     ,            usd_amount
     ,       EXTRACT (YEAR FROM invoice_date)
,     got_neighbor_values     AS
     SELECT     id
     ,     invoice_date
     ,     usd_amount
     ,     this_yr
     ,     LAG (this_year)  OVER ( PARTITION BY  id
                                ,           usd_amount
                            ORDER BY      this_year
                          )           AS prev_year     
     ,     LEAD(this_year)  OVER ( PARTITION BY  id
                                ,           usd_amount
                            ORDER BY      this_year
                          )           AS next_year
     FROM     got_this_yr
SELECT       n.id
,       c.name
,       c.product
,       c.start_delivery
,       c.end_delivery
,       n.invoice_date
,       n.usd_amount
FROM       got_neighbor_values     n
JOIN       customer          c  ON     c.id     = n.id
WHERE       n.this_year     IN ( n.prev_year + 1
                    , n.next_year - 1
ORDER BY  n.id
,            n.usd_amount
,       n.invoice_date
;If you'd care to post CREATE TABLE and INSERT statements for the sample data, then I could test this.
This used the analytic LAG and LEAD functions to find the previous and next years when the same amount occurred with the same id. Since we want to use this value in a WHERE clause, it has to be computed in a sub-query (got_neighbor_vlues) first; analytic functions are always computed after the WHERRE clause is applied.
In the case of the very first year, LAG will return NULL, and in the case of the very last year, LEAD will return NULL. That's okay for our purposes; we just need something such that the WHERE clause of the main query will be TRUE if and only if one of the adjacent rows was a consecutive year.
Why don't you want a row for 2007 in the results? It looks like the same customer had the exact same amount in 2007, 2008 and 2009.
What happens if the same customer has the exact same amount 2 (or more) times in the same calendar year? The query above uses the earliest one in each year, and ignores the others. If that's not what you want, then post the results you do want
Edited by: Frank Kulash on Nov 15, 2010 10:36 AM
Included cutomer table