|
|
|
SSC Veteran
      
Group: General Forum Members
Last Login: Thursday, June 06, 2013 9:57 AM
Points: 234,
Visits: 623
|
|
Hi all - I have an issue I've been fighting with and can't seem to come up with a solution. Basically here is the pseudo code:
For each RecordID in tmpEvalRecs I need to see if any other order records in tmpOrders have duplication based on ProductID, OrderDate, ProductClass, and ProductType.
1. Gather the recordid from tmpEvalRecs 2. Gather ProductID, OrderDate, ProductClass, and ProductType from tmpOrders where recordid = recordid from above 3. Look for any other orders that have the same line items based on the fields above, in the same order (by orderdate). 4. I need the original recordid and duplicate RecordID returned
Any assistance would be greatly appreciated. I'm trying to avoud itterating through the records one by one.
Here is some table scripts and data: CREATE TABLE tmpEvalRecs ( RecordID INT, ); CREATE TABLE tmpOrders ( RecordID INT, ProductID VARCHAR(10), OrderDate DATETIME, ProductClass CHAR(1), ProductType VARCHAR(20), ProductRep VARCHAR(20), ReferralNumber INT );
INSERT INTO tmpEvalRecs SELECT 1 UNION SELECT 3 UNION SELECT 4;
INSERT INTO tmpOrders SELECT 1, 'PD101', '07/01/2009', 'C', 'WidgetA', 'Sally', 21 UNION SELECT 1, 'PD102', '07/02/2009', 'B', 'WidgetB', 'Sally', 21 UNION SELECT 1, 'PD103', '07/03/2009', 'A', 'WidgetC', 'Sally', 21 UNION SELECT 2, 'PD101', '07/01/2009', 'C', 'WidgetA', 'Sally', 21 UNION SELECT 2, 'PD102', '07/02/2009', 'B', 'WidgetB', 'Sally', 21 UNION SELECT 2, 'PD103', '07/03/2009', 'A', 'WidgetC', 'Sally', 21 UNION SELECT 3, 'PD101', '07/01/2009', 'C', 'WidgetA', 'Sally', 21 UNION SELECT 3, 'PD105', '07/08/2009', 'B', 'WidgetB', 'Sally', 21 UNION SELECT 4, 'PD101', '07/01/2009', 'C', 'WidgetA', 'Sally', 21 UNION SELECT 5, 'PD101', '07/01/2009', 'C', 'WidgetA', 'Sally', 21;
|
|
|
|
|
SSC Veteran
      
Group: General Forum Members
Last Login: Thursday, May 09, 2013 4:22 PM
Points: 285,
Visits: 1,575
|
|
What is your expected output?
See if this works for you:SELECT DISTINCT O2.* FROM tmpOrders AS O1 INNER JOIN tmpOrders AS O2 ON O1.RecordID <> O2.RecordID AND O1.ProductID = O2.ProductID AND O1.OrderDate = O2.OrderDate AND O1.ProductClass = O2.ProductClass AND O1.ProductType = O2.ProductType LEFT OUTER JOIN tmpEvalRecs AS Rec ON O2.RecordID = Rec.RecordID WHERE Rec.RecordID IS NULL
|
|
|
|
|
SSC-Enthusiastic
      
Group: General Forum Members
Last Login: Wednesday, April 20, 2011 6:49 AM
Points: 129,
Visits: 143
|
|
;WITH CTEs AS ( SELECT O.ProductID, O.OrderDate, O.ProductClass, O.ProductType, DENSE_RANK() OVER (PARTITION BY O.ProductID,O.OrderDate,O.ProductClass,O.ProductType ORDER BY NEWID()) 'Duplicate' FROM tmpEvalRecs R(NOLOCK) JOIN tmpOrders O(NOLOCK) ON(R.recordid = O.recordid) )
SELECT * FROM CTEs WHERE Duplicate > 1
Note: = 1 (Single Occurance) >1 (Multiple Occurance)
Cheers
|
|
|
|
|
Old Hand
      
Group: General Forum Members
Last Login: Yesterday @ 2:40 AM
Points: 356,
Visits: 1,667
|
|
If the you looking for Duplicates based on productids then this should work:
select COUNT(tmpOrders.productid) as duplicates,tmpOrders.productid,tmpOrders.productclass,tmpOrders.OrderDate,tmpOrders.ProductType from tmpOrders group by tmpOrders.productid,tmpOrders.productclass,tmpOrders.OrderDate,tmpOrders.ProductType having COUNT(tmpOrders.productid)>1
blog: http://sarveshsingh.com
Twitter: @sarveshsing
|
|
|
|
|
Old Hand
      
Group: General Forum Members
Last Login: Yesterday @ 2:40 AM
Points: 356,
Visits: 1,667
|
|
i think u needed the recordid as well,,,
with CTE as ( select *, ROW_NUMBER() over (PARTITION by productid,productclass,producttype,orderdate order by orderdate) as rn from tmpOrders ) select * from CTE where rn>1;
blog: http://sarveshsingh.com
Twitter: @sarveshsing
|
|
|
|
|
SSC Veteran
      
Group: General Forum Members
Last Login: Thursday, June 06, 2013 9:57 AM
Points: 234,
Visits: 623
|
|
First - Thank you to all that spent time working on a solution.
Lamprey13's solution is hitting on the correct records. If you look at the tmpOrders data sorted by recordid and orderdate, recordid 2 is a match for recordid 1 based on ProductID,OrderDate,ProductClass, and ProductType. Also, recordid 5 is a match for recordid 4 based on the same fields sorted by orderdate.
My next challenge is to include both the matched data (as the query is doing now) and the data from the recordid being evaluated.
Ex: If recordid 2 is found as a match to recordid 1, return the order details for both record id's. Is this possible?
|
|
|
|