执行pandas groupby *** 作的更快选择

执行pandas groupby *** 作的更快选择 脾气暴躁的

numpy.add.at

和

pandas.factorize

目的是要快。但是，我也尝试将其组织为易于阅读。

i, r = pd.factorize(df.name)j, c = pd.factorize(df.color)n, m = len(r), len(c)b = np.zeros((n, m), dtype=np.int64)np.add.at(b, (i, j), 1)pd.Series(c[b.argmax(1)], r)John     WhiteTom       BlueJerry    Blackdtype: object

---

groupby

，

size

和

idxmax

df.groupby(['name', 'color']).size().unstack().idxmax(1)nameJerry    BlackJohn     WhiteTom       Bluedtype: objectnameJerry    BlackJohn     WhiteTom       BlueName: color, dtype: object

---

Counter

¯_(ツ)_/¯

from collections import Counterdf.groupby('name').color.apply(lambda c: Counter(c).most_common(1)[0][0])nameJerry    BlackJohn     WhiteTom       BlueName: color, dtype: object