Comment by Quothling

> People usually approach this the other way, use something like pandas or numpy from the beginning if it solves your problem.

That is exactly how we approach it though. We didn't start out with turbodbc + pandas, it started as an sql alchemy and pandas service. Then when it was too slow, I got involved, found and dealth with the bottle necks. I'm not sure how you would find and fix such things without knowing the efficiency or lack there of in different parts of Python. Also, as you'll notice, we didn't write ur own stuff, we simply used more efficient Python libraries.