Tracking changes made to an address block of info in a database

David Thompson dthmpsn1 at uiuc.edu
Wed Mar 5 15:13:56 PST 2008


>Dave, how about a character count approach where I add up the number of
>characters of the block in question and assign it to another field and that
>way a tab through the block wouldn't change count and not trigger a change
>in timestamp. Possibly doing something like if char count of block not =
>original value in tracking field then set current date/time/user in the
>tracking field?

That doesn't sound like a very reliable way of detecting a change. If 
someone moves to another address on the same street, there wouldn't 
be any change in the character count. You would also have cases where 
the street name got longer but the city name got shorter and they 
cancelled, and so on. I'll bet a field with that kind of character 
count would have a high percentage of the counts clustered about some 
middle value. With so many different addresses sharing the same 
counts there are sure to be cases where the address changes, and the 
count doesn't. Josh's suggestion wouldn't require any extra memory to 
speak of. If that's the concern, that's probably the way you would 
want to go.

Dave


More information about the Qna mailing list