You've got a classic problem there, Martin, and there's no easy
solution. The way this is usually approached is to first standardize
your street names in both tables. You'll have to come up with ways to
identify when a street name doesn't meet your standards, and then the
second task is to correct it. I'd also recommend that use new columns
for normalizing your streets so you'll always have the original. Also,
you might want to do something non-standard, like sort all the big words
in alphabetical order to catch the "Kosciuszki Tadeusza" situations you
mentioned above.
MapInfo's MapMarker can be used to normalize addresses, and does a very
good job at this, but it'll cost you. If you want save some money and
spend some time instead, there are Open Source solutions. The best one
is SRC's Explorer Geocoder. It can work with MapBaaic too through a
DLL-based API.
I don't have the download info handy, but I'm sure Google knows. The
best search terms I can come up with are "geocoding" and "address
normalization".
I think it's actually less work cleaning your data first than trying to
SQL your way around all the many variations in databases that capture
words and names of people and places.
Good luck!
--
- Bill Thoen
GISnet - www.gisnet.com
303-786-9961