text matching

[open.network.design]

I have a database, see attached, that has two tables one table called "inventory" for items that we have in our inventory, currently about 400,000 records. The other table called "bad" has items that we want to remove from our inventory. The only field that we can use is the items name but as the item names are just a bit different in each table there are no exact matches. If you see our current setup there are about 1/5 of the items it catches but the rest are just not exactly the same. Is there any way to pattern match the words from one table to another and based on a percentage of matched words or make some type of way to remove these few hundred items from the larger inventory table? Its does not have to be perfect just good enough to catch most of the items from the bad table. Any help would be great!

Thanks

sample.fp7.zip

[comment]

I would start by removing common words such as "a", "an", "the", "on", "of", "at", etc. That would extend the match to records that have all significant words in common, in the same order.

The next step up could be determining the percentage of matched words - but that would be quite difficult to do in version 8 (and slow in any version).