On the contrary, I like the suggestion of not needing to replicate all data. That is also a good point of encountering errors during a large process.

I am thinking of making a generic function that when given a table/object name, would be flexable enough to pull all the data from that table and put it into mysql. This function would be passed an array or looped through one so that each part would be seperate and could log errors or repeat a single section if errors were encountered. I want to keep it modular, I just don't want to write a different replication script for each table. I will work on this and let you know how it goes.