I'm working on a project at work that is a re-write of a piece of functionality from ASP into .NET.
So the gist in a nutshell is that the new .NET forms should function exactly the same way as the ASP forms do. To the point where saving these records into the database should match excluding things like primary keys & insert date/time stamps.
To throw an additional wrench into the mix there are two DBs these records get saved out into. The first datasource is as denormalized as you can get. And the second uses a Sync PROC to move things from the first into the second and normalize the living daylights out of it.
So basically what you end up with is 8 key tables in the first datasource and over 20 in the second.
What I have been doing is running a sequence of queries for a legacy data row and .NET data row where the data entered on the forms themselves was identical excluding the unique identifier fields (email & SSN for example).
So the queries end up looking something like
select * from table1 where ID=LegacyID
select * from table1 where id=.NETID
Then I compare the results.
Now here's the kicker--EVERYTIME I do this I think to myself God there must be a better way...and then I think, Nah, there's no point in finding a better solution I won't have to do this comparison again.
I'm up to round 3 at this point. Thankfully, the developers and I have an agreement. I'm not going to progress to the normalized DB until the de-normalized datasource is acceptable, which for right now its not. *WHEW*
Pondering whether delivering an "automated" method would be better or not...
I guess if I have to do this one or two more times, I'm gonna opt for an automated solution. Hell I could probably write a QTP script that would do this..I just haven't yet. NOTE--I stress YET here.
*grumbles*