Hi Justin! > By testing for (shared house), you'd end up grouping them > together like > this: > > Fred Flintstone > Wilma Flintstone > Homer Simpson > Marge Simpson > Bedrock > (shared house) > > ... which isn't your intended result.
Yes it is ;))))) What you write is right but this is how I did it! I've indeed got one medics table and one for practice/premise. So in $medTable I simply store the id of the premise a doctor is working. > > $queryString .= "FROM $medTable m, $praxTable p WHERE "; But I read from both tables, so I got the description and the id; > Really, what you'd like to do is group people together based on their > premises being the same. This of course is complicated by > the fact that if > the premises is misspelled, or changed slighly (Unit 2, Suite > The real solution is a change in your data design. What you > actually want > is to relate Doctors to their premises/practice. Hence, I > would have a > table of Doctors, and a table of Premises, with an id being > the key for > each. > This way Doctor #2 and Doctor #15 can both be associated with > Premises #5. > Better still, a Doctor may be associated with one or more practices. > This seems like a smarter/quicker/faster/more future proof method than > hoping for matches in the address and performing complex array sorts. What you advice me to do is IMO the one and only right way and is part of a standard normalisation process, isn't it? Nevertheless, Thanx for your hint; If I not already had this this would be an important one! And btw., I learnd a new english word, too (premise). So I always keep learning ;) Cheers, Kiko -- It's not a bug, it's a feature. christoph starkmann -- -- PHP General Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php
