Matching like, but not exact data

[Craig]

Can someone point me to documented techniques used to
match similar data? For example, I have two tables of
contacts:

Table 1
Name: Joe Smith
Street1: 123 Main St
Street 2: P.O. Box 123

Table 2
Name: Joseph Smith
Street 1: PO Box 123
Street 2: null

You see my challenge. Obviously the record in each table
is the same contact, but I have a couple issues when
trying to match. The data can be different (Joe vs.
Joseph, P.O. vs. PO) and/or the similar data can be in
different fields. Each table has thousands of records,
so it would be a tedious process, and not fool proof to
match on partial data and eyeball. Any ideas?

Thanks for taking the time.
Craig

 

[John Vinson]

Each table has thousands of records,
so it would be a tedious process, and not fool proof to
match on partial data and eyeball.

It would be tedious indeed. This is why the numerous address-list
cleanup services charge breathtaking fees for doing so.

There *is* no simple automated solution; it requires human
intelligence to cover all the multifarous possibilities. I wish I
could suggest one, but AFAIK artificial intelligence simply hasn't
advanced to that point!