Issue: New Import after Interrupted Import results in many problems

Continuing from this issue: http://www.vaultwiki.org/issues/1604/

I attempted the secondary import to fill in the gaps from my first apparently partial import. After doing so I had the following problems:
1- Every page in the wiki now had a duplicate entry
2- All changes previously made were erased and replaced by the original broken imported pages
3- Attempting to rename one of the duplicate pages to see if the other was my newer edited version had no result, as either both copies were the same or it was forcing a redirect to the renamed page.

I was forced to do a fast restore of the MySQL database to my automatic back-up from a day before, and forgot to check to see whether or not the second import managed to import the pages that the first import missed.

My suggestion for this issue would be to create some kind of tag or code attached to imported pages based off of the original mediawiki page's title, and then have the importer ignore importing any pages with a tag matching a tag already existing in the database. Then you could even break the importer script down into smaller repeatable chunks and just let it cycle through until an import was completely finished. That just my thought based on how I remember the Impex importer working, not sure how practical it is.