Well now he has 2 solutions... his choice to make. But for the sake of future requirements I'd add the releasedate column and at least try to "normalize" the...

I totally agree with ron... it's better to have less data that is accurate than all the data that might be...

1 - Not sure, check books online

2 - Just create a datetime column, then create a trigger that will update that column everytime the row is modified.

3 - It can...

The best solution would be to create a new ReleaseDate column (assuming you can't delete/replace the issue column). All you'd have to do is figure out the first release...

Not really because you have 52 weeks in a year and 12 months times 4 = 48.

Can you show us some sample data so we can find a work around?

If you had a date column you could do group by month(date).

1 - timestamp is not a date (can't use date functions on this)

2 - timestamp is maintained by sqlserver on the insert and each update whereas the datetime must be...

You'll need to post the table definition along with some sample data so we can help you with your query.

(num.number + 1.0)/@NumPayments

Or declare @NumPayments as decimal if it's possible... but I'd got with + 1.0

I'll leave this one in the hands of the gurus... I don't know any other ways.

You'll have to put a trigger on that table, and when the field(s) of the adress are changed, you'll have to insert the old data in another table to keep...

You either have to set a trace on the server which can cause a lot of overhead...

Or you can make sure that the users can only access the data from...

R u doing this in a DTS?

If you can't work it out, post your query and we'll help you.

This will need some tweaking but it can give you an idea :

SELECT [city_code], [outlet_code], sum([sold_units]) as UnitsSold, CAST(Issue - 1 AS INT) / 4 AS UnitPeriod

FROM [MerchDB].[dbo].[tblOutletSales]

where active...