I have a legacy database (actually a single fully de-normalized table with MANY columns) that I need to move over to SQL and a more orderly data structure. However, much of the data was entered as free-text, so I have columns of things such as names, cities, etc. that vary slightly depending on who did the entering. In the simplest cases, extra white space or different capitalization are used. In more complex cases, nouns are spelled differently or even a word (like a middle name) is left out. So, I need to somehow consolidate these textual variations describing the same thing into a single entity. Concrete example:

Promessa High Spirits
Promessa high Spirits
Promessa Hgh Spirits
Promessa high spirits

These all represent the same animal. I would like to "discover" that (or at least be alerted to the situation) without having to go through the 50000 records like this that I have.

Thanks,
Sean