Tolerant filter

[Gottfried Lesigang]

Hi NG!

At the moment I'm in charge of "cleaning". I "clean" the db of a customer
and find many datas (mostly names and addresses) twice - only distinguished
by a small difference, maybe even a spelling mistake. As these lines aren't
identical I can't simply check this.

Well, I'm done with the old data very soon, but in future I want to prevent
users from inputting a "new" person without checking first whether it exists
or not.

The "normal" user will not be able to type in a criterium for a
SQL-"LIKE"-Statement (such as "*ma??er*")

Which ways do you know to perform a search of "similarity" (preferably in
german language)? It's result should be all lines which are somehow similar
to the one I'm looking for.

Regards and TIA
Gottfried

 

[Marshall Barton]

At the moment I'm in charge of "cleaning". I "clean" the db of a customer
and find many datas (mostly names and addresses) twice - only distinguished
by a small difference, maybe even a spelling mistake. As these lines aren't
identical I can't simply check this.

Well, I'm done with the old data very soon, but in future I want to prevent
users from inputting a "new" person without checking first whether it exists
or not.

The "normal" user will not be able to type in a criterium for a
SQL-"LIKE"-Statement (such as "*ma??er*")

Which ways do you know to perform a search of "similarity" (preferably in
german language)? It's result should be all lines which are somehow similar
to the one I'm looking for.

Try searching the web for SoundX algorithms.