After you use the find similar command in the other tab in the tree there is a duplicates icon on the left you can open in a new window and delete from there. By default find similar is a read only result as a safety precaution. Duplicate database records are generally not a big deal as they take up very little space unless you have a inordinate amount or make a lot of thumbnails. I would rather get too many results of a query than none at all. Also with the copy and paste of metadata you can merge the metadata to ensure your records match! You may also want to try Media file as it looks for entries with the same path.