数据清洗之 分组对象与apply函数

Posted 2022-07-08 wx62c62b36cedf9

分组对象与apply函数

• 函数apply即可用于分组对象，也可以作用于dataframe数据
• Groupby.apply(func)
• 需要注意axis=0和axis=1的区别

• np.sum, axis=0 相当于计算每列的总和
• np.sum, axis=1 相当于计算每行的总和

import pandas as pd
import numpy as np
import

os.getcwd()

D:\\\\Jupyter\\\\notebook\\\\Python数据清洗实战\\\\数据清洗之数据统计

os.chdir(D:\\\\Jupyter\\\\notebook\\\\Python数据清洗实战\\\\数据)

df = pd.read_csv(online_order.csv, encoding=gbk, dtype=customer:str, order:str)

df.head(5)

	customer	order	total_items	discount%	weekday	hour	Food%	Fresh%	Drinks%	Home%	Beauty%	Health%	Baby%	Pets%
0	0	0	45	23.03	4	13	9.46	87.06	3.48	0.00	0.00	0.00	0.0	0.0
1	0	1	38	1.22	5	13	15.87	75.80	6.22	2.12	0.00	0.00	0.0	0.0
2	0	2	51	18.08	4	13	16.88	56.75	3.37	16.48	6.53	0.00	0.0	0.0
3	1	3	57	16.51	1	12	28.81	35.99	11.78	4.62	2.87	15.92	0.0	0.0
4	1	4	53	18.31	2	11	24.13	60.38	7.78	7.72	0.00	0.00	0.0	0.0

grouped = df.groupby(weekday)

# 只可传入一个统计参数
# agg可传入多个
# grouped.apply([np.mean, np.sum])

grouped.apply(np.mean)[[total_items, discount%, weekday]]

	total_items	discount%	weekday
weekday
1	30.662177	8.580705	1.0
2	31.868612	8.638014	2.0
3	31.869796	7.794507	3.0
4	32.251899	8.068155	4.0
5	31.406619	9.159031	5.0
6	32.154814	8.414258	6.0
7	32.373837	8.710171	7.0

df.columns

Index([customer, order, total_items, discount%, weekday, hour,
       Food%, Fresh%, Drinks%, Home%, Beauty%, Health%, Baby%,
       Pets%],
      dtype=object)

var = [Food%, Fresh%, Drinks%, Home%, Beauty%, Health%, Baby%,
       Pets%]

df[var].head(5)

	Food%	Fresh%	Drinks%	Home%	Beauty%	Health%	Baby%	Pets%
0	9.46	87.06	3.48	0.00	0.00	0.00	0.0	0.0
1	15.87	75.80	6.22	2.12	0.00	0.00	0.0	0.0
2	16.88	56.75	3.37	16.48	6.53	0.00	0.0	0.0
3	28.81	35.99	11.78	4.62	2.87	15.92	0.0	0.0
4	24.13	60.38	7.78	7.72	0.00	0.00	0.0	0.0

# 计算每个变量的总和
df[var].apply(np.sum, axis=0)

Food%      706812.19
Fresh%     606818.38
Drinks%    700477.06
Home%      406187.25
Beauty%    176788.48
Health%     33988.76
Baby%      332884.34
Pets%       31292.61
dtype: float64

# 对每一行求和
df[var].apply(np.sum, axis=1).head(5)

0    100.00
1    100.01
2    100.01
3     99.99
4    100.01
dtype: float64

# Food% - Fresh%
df[var].apply(lambda x: x[0] - x[1], axis=1).head(5)

0   -77.60
1   -59.93
2   -39.87
3    -7.18
4   -36.25
dtype: float64