Question

I am not looking for help with the actual code of my stored procedure (yet anyway), but a client sent me a copy of his SQL Server 2000 database stating that a particular proc was returning incorrect results; I restored the database to my server (SQL Server 2008R2) and ran the proc and it produced the correct results. Keep in mind there is no front-end that could be causing the difference – in both the cases I simply execute the proc thru enterprise manager and management studio.

I am wondering if there are any known differences that might contribute/cause this problem? This particular proc uses a lot of conditional logic, default parameters and even has a cursor to do some looping logic, but any suggestions on possible areas to look at would be helpful.

Edit: to clarify, the number of rows and columns returned are the same, but for a particular column (for certain rows depending on value of a data field), the number is exactly double what it should be. 2008 gives the ‘correct’ answer. ‘2000’ gives the wrong answer, i.e. the double amount.

Answer 1

There are so many things that could cause this problem including (but not limited to) collation and system settings. If we saw the SP we might be able to spot the culprit. In addition to changes in deployments there are differences between versions see

http://msdn.microsoft.com/en-us/library/ms143179(v=sql.100).aspx

and

http://msdn.microsoft.com/en-us/library/ms143179.aspx