Are for-loops in pandas really bad? When should I care?

Question

Are for loops really \"bad\"? If not, in what situation(s) would they be better than using a more conventional \"vectorized\" approach?¹

Answer 1

In short

• for loop + iterrows is extremely slow. Overhead isn't significant on ~1k rows, but noticeable on 10k+ rows.
• for loop + itertuples is much faster than iterrows or apply.
• vectorization is usually much faster than itertuples

Benchmark