PERF: Join non unique by lukemanley · Pull Request #56817 · pandas-dev/pandas (original) (raw)

Seeing ~10% improvement when how is "left", "right", or "inner".

import pandas as pd 
import numpy as np

idx1 = pd.Index(np.tile(np.arange(1000), 1000))
idx2 = pd.Index(np.arange(100))

%timeit idx1.join(idx2, how="left")

# 103 ms ± 1.61 ms per loop (mean ± std. dev. of 7 runs, 10 loops each)  -> main
# 90.6 ms ± 504 µs per loop (mean ± std. dev. of 7 runs, 10 loops each)  -> PR