Why is Spark running slower than Python with large files?

Question

I know that one of the main reasons to use Apache Spark is to make things faster than doing it "manually" with a Python function.

However, I have this file call