pandas——常用操作

常用操作

import pandas as pd
import numpy as np

example = pd.DataFrame({'Amount': [74., 235., 175., 100., 115., 245., 180., 90., 88., 129., 273., 300.],
              'Category': ['Transportation', 'Grocery', 'Household', 'Entertainment', 
                           'Transportation', 'Grocery', 'Household', 'Entertainment', 
                           'Transportation', 'Grocery', 'Household', 'Entertainment'],
              'Month': ['January', 'January', 'January', 'January', 
                        'February', 'February', 'February', 'February', 
                        'Marth', 'Marth', 'Marth', 'Marth']})

example

	Amount	Category	Month
0	74.0	Transportation	January
1	235.0	Grocery	January
2	175.0	Household	January
3	100.0	Entertainment	January
4	115.0	Transportation	February
5	245.0	Grocery	February
6	180.0	Household	February
7	90.0	Entertainment	February
8	88.0	Transportation	Marth
9	129.0	Grocery	Marth
10	273.0	Household	Marth
11	300.0	Entertainment	Marth

1. 排序操作

example.sort_values(by=['Month', 'Amount'], ascending=[True, False])   # by 指定按照哪些属性排序 Month放在第一个 先把Month 按照升序排序，再对Amount降序排序

	Amount	Category	Month
5	245.0	Grocery	February
6	180.0	Household	February
4	115.0	Transportation	February
7	90.0	Entertainment	February
1	235.0	Grocery	January
2	175.0	Household	January
3	100.0	Entertainment	January
0	74.0	Transportation	January
11	300.0	Entertainment	Marth
10	273.0	Household	Marth
9	129.0	Grocery	Marth
8	88.0	Transportation	Marth

data = pd.DataFrame({'k1':[1,2,3,5,5,4,6,2,2], 'k2':[12,8,11,4,4,52,30,18,18]})
data

	k1	k2
0	1	12
1	2	8
2	3	11
3	5	4
4	5	4
5	4	52
6	6	30
7	2	18
8	2	18

data.sort_values(by='k2')

	k1	k2
3	5	4
4	5	4
1	2	8
2	3	11
0	1	12
7	2	18
8	2	18
6	6	30
5	4	52

data.drop_duplicates()    # 去除完全一样的数据

	k1	k2
0	1	12
1	2	8
2	3	11
3	5	4
5	4	52
6	6	30
7	2	18

data.drop_duplicates(subset=['k1'])    # 只要 k1 属性的值一样 就去除

	k1	k2
0	1	12
1	2	8
2	3	11
3	5	4
5	4	52
6	6	30

data2 = pd.DataFrame({'k1':[1,5,3,16,19], 'k2':['A1','A2', 'A1', 'B2', 'A2']})
data2

	k1	k2
0	1	A1
1	5	A2
2	3	A1
3	16	B2
4	19	A2

如果希望将 A1和A2 都归为A类

def map(series):
    if series['k2'] == 'A1':
        return 'A'
    elif series['k2'] == 'A2':
        return 'A'

data2['k2_map'] = data2.apply(map, axis=1)
data2

	k1	k2	k2_map
0	1	A1	A
1	5	A2	A
2	3	A1	A
3	16	B2	None
4	19	A2	A

df = pd.DataFrame({'t1':np.random.randn(5), 't2':np.random.randn(5)})
df

	t1	t2
0	0.909904	-0.394096
1	0.943165	0.480281
2	-1.616222	2.153146
3	-0.403424	-1.160060
4	0.850074	1.280206

df2 = df.assign(ration = df['t1'] / df['t2'])    # 新插入一列属性 ration 值为 t1的值/t2的值

df2

	t1	t2	ration
0	0.909904	-0.394096	-2.308840
1	0.943165	0.480281	1.963779
2	-1.616222	2.153146	-0.750633
3	-0.403424	-1.160060	0.347762
4	0.850074	1.280206	0.664013

df2.drop('ration', axis=1)     # 删除 ration 这一列

	t1	t2
0	0.909904	-0.394096
1	0.943165	0.480281
2	-1.616222	2.153146
3	-0.403424	-1.160060
4	0.850074	1.280206

将数据按指定区间切分

age = pd.Series([18, 9, 6, 20, 26, 67, 52, 38, 40, 81, 36, 74, 88])
bins = [10, 40, 60, 90]     # 将数据切分为 10-40 40-60 60-90 的区间
res = pd.cut(age, bins)
res

0     (10.0, 40.0]
1              NaN
2              NaN
3     (10.0, 40.0]
4     (10.0, 40.0]
5     (60.0, 90.0]
6     (40.0, 60.0]
7     (10.0, 40.0]
8     (10.0, 40.0]
9     (60.0, 90.0]
10    (10.0, 40.0]
11    (60.0, 90.0]
12    (60.0, 90.0]
dtype: category
Categories (3, interval[int64, right]): [(10, 40] < (40, 60] < (60, 90]]

有的数据不在区间范围内就被设置为 NaN

pd.value_counts(res)     # 对每个区间进行计数

(10, 40]    6
(60, 90]    4
(40, 60]    1
dtype: int64

w = pd.cut(age, bins, labels=['youth', 'mid', 'old'])     # 给每个区间设置一个名称 数量应与区间数一致
pd.value_counts(w)

youth    6
old      4
mid      1
dtype: int64

res.isnull()      # isnull() 方法判断每一个值是否为NAN   DataFrame也是一样的方法

0     False
1      True
2      True
3     False
4     False
5     False
6     False
7     False
8     False
9     False
10    False
11    False
12    False
dtype: bool

qw = pd.Series([np.nan, 1, 2, 3])  
qw

0    NaN
1    1.0
2    2.0
3    3.0
dtype: float64

qw.fillna(5)   # 用一个指定值 填充所有 NaN

0    5.0
1    1.0
2    2.0
3    3.0
dtype: float64