df:
A
0 219
1 590
2 272
3 945
4 175
5 930
6 662
7 472
8 251
9 130
我正在尝试基于值所属的分位数创建新的列分位数,例如:
if value > 1st quantile : value = 1
if value > 2nd quantile : value = 2
if value > 3rd quantile : value = 3
if value > 4th quantile : value = 4
代码:
f_q = df['A'] .quantile (0.25)
s_q = df['A'] .quantile (0.5)
t_q = df['A'] .quantile (0.75)
fo_q = df['A'] .quantile (1)
index = 0
for i in range(len(test_df)):
value = df.at[index,"A"]
if value > 0 and value <= f_q:
df.at[index,"A"] = 1
elif value > f_q and value <= s_q:
df.at[index,"A"] = 2
elif value > s_q and value <= t_q:
df.at[index,"A"] = 3
elif value > t_q and value <= fo_q:
df.at[index,"A"] = 4
index += 1
代码工作正常。但是我想知道是否有更有效的方法来做到这一点。任何建议都是有帮助的。
答案 0 :(得分:2)
是,使用pd.qcut
:
>>> pd.qcut(df.A, 4).cat.codes + 1
0 1
1 3
2 2
3 4
4 1
5 4
6 4
7 3
8 2
9 1
dtype: int8
(给我与您的代码完全相同的结果。)
您还可以在qcut
结果上致电np.unique
:
>>> np.unique(pd.qcut(df.A, 4), return_inverse=True)[1] + 1
array([1, 3, 2, 4, 1, 4, 4, 3, 2, 1])
或者,使用pd.factorize
(请注意输出中的细微差别):
>>> pd.factorize(pd.qcut(df.A, 4))[0] + 1
array([1, 2, 3, 4, 1, 4, 4, 2, 3, 1])