参考文章：https://jizhi.im/blog/post/10min2pandas01
https://jizhi.im/blog/post/10min2pandas02
导入pandas和numpy库
import pandas as pd
import numpy as np
一、创建对象
1、创建Series对象
创建Series对象，需要传递一个List参数，
# Series，可以指定index，同样是传递List参数
s = pd.Series([1,3,5,np.nan,6,8],index=[i+1 for i in range(len([1,3,5,np.nan,6,8]))])
print(s)
#输出
'''
1 1.0
2 3.0
3 5.0
4 NaN
5 6.0
6 8.0
dtype: float64
'''
2、创建Dataframe对象
# 首先创建一个时间序列
dates = pd.date_range('20130101', periods=6,freq='d')
'''
参数：
:param start: 序列开始时间
:param end: 序列结束时间, 给定start时, 结束时间不包含end
:param periods: int, 生成的时间序列长度
:param freq: 要生成时间序列的时间间隔，可以是天（d）、多天（2d）、月（M）、多月（3M）、年（Y）、多年（2Y）
以及时（h）分（min）秒（s）等
:param out_format: 是否输出格式化后的字符串, 若要输出可指定输出格式. "%Y-%m-%d %H:%M:%S"
:param input_format: 若start或end是字符串且无法自动推断时间格式则需指定格式
'''
print(dates)
# 输出
'''
DatetimeIndex(['2013-01-01', '2013-01-02', '2013-01-03', '2013-01-04',
'2013-01-05', '2013-01-06'],
dtype='datetime64[ns]', freq='D'
'''
# 再创建一个6行4列的随机数组
data = np.random.randn(6,4)
print(data)
# 输出
'''
[[ 0.52615987 -0.10511352 0.93578156 -1.0171928 ]
[ 0.14465037 0.92757347 0.31852467 -0.82320396]
[-0.63506826 0.62381754 -0.61869088 -0.23659419]
[ 0.53514965 0.17191194 -1.25918251 -2.80756598]
[-0.97158469 0.51641578 0.71967892 0.48177381]
[-0.82879759 1.21244248 0.21248849 -0.62200675]]
'''
# 最后创建DataFrame，index指定索引，columns通过传递List指定列名
df = pd.DataFrame(data,index=dates,columns=['A','B','C','D'])
print(df)
#输出
'''
A B C D
2013-01-01 0.526160 -0.105114 0.935782 -1.017193
2013-01-02 0.144650 0.927573 0.318525 -0.823204
2013-01-03 -0.635068 0.623818 -0.618691 -0.236594
2013-01-04 0.535150 0.171912 -1.259183 -2.807566
2013-01-05 -0.971585 0.516416 0.719679 0.481774
2013-01-06 -0.828798 1.212442 0.212488 -0.622007
'''
# 也可以通过传递字典构建DataFrame
dict = { 'A' : 1., 'B' : pd.Timestamp('20130102'), 'C' : pd.Series(1,index=list(range(4)),dtype='float32'),
'D' : np.array([3] * 4,dtype='int32'), 'E' : pd.Categorical(["test","train","test","train"]), 'F' : 'foo' }
print(dict)
# 输出
{'A': 1.0, 'B': Timestamp('2013-01-02 00:00:00'), 'C': 0 1.0
1 1.0
2 1.0
3 1.0
dtype: float32, 'D': array([3, 3, 3, 3]), 'E': [test, train, test, train]
Categories (2, object): [test, train], 'F': 'foo'}
df2 = pd.DataFrame(dict)
print(df2)
print('=' * 60)
print(df2.dtypes)
# 输出
A B C D E F
0 1.0 2013-01-02 1.0 3 test foo
1 1.0 2013-01-02 1.0 3 train foo
2 1.0 2013-01-02 1.0 3 test foo
3 1.0 2013-01-02 1.0 3 train foo
============================================================
A float64
B datetime64[ns]
C float32
D int32
E category
F object
dtype: object
二、观察数据
# 默认情况下，.head()和.tail()输出首尾的前5行，也可以手动指定输出行数。
df.head()
df.tail(n=5)
# 索引，如果不添加.values，返回的数据类型是index，否则返回的数据类型是数组
print(df.index)
print(type(df.index))
# 输出
'''
Index(['A', 'B', 'C', 'D'], dtype='object')
<class 'pandas.core.indexes.base.Index'
'''
print(df.index.values)
print(type(df.index.values))
# 输出
'''
['2013-01-01T00:00:00.000000000' '2013-01-02T00:00:00.000000000'
'2013-01-03T00:00:00.000000000' '2013-01-04T00:00:00.000000000'
'2013-01-05T00:00:00.000000000' '2013-01-06T00:00:00.000000000']
<class 'numpy.ndarray'>
'''
# 列标签
print(df.columns)
print(df.columns.values)
# 数值，返回数组
df.values
# 描述统计
dt.describe() # 默认只统计数值类型的特征
dt.describe(include='all') # 全统计
dt.describe(include=[np.number]) # 只统计数值类型的特征
dt.describe(include=[np.object]) # 只统计字符类型的特征
dt.describe(include=['列名']) # 统计指定列
dt.describe(exclude=[np.number]) # 反向指定
三、转置与排序
1、转置
print(df.T)
2、排序
print(df2.sort_index(axis=1, ascending=False)) # 按轴排序，逐列递减
print('=' * 60)
print(df2.sort_values(by='E', ascending=False)) # 按值排序，'E'列逐行递增
print('=' * 60)
print(df2[['A','C','E']].sort_values(by=['E', 'C'], ascending=False)) # by可以传递List
# 输出
F E D C B A
0 foo test 3 1.0 2013-01-02 1.0
1 foo train 3 1.0 2013-01-02 1.0
2 foo test 3 1.0 2013-01-02 1.0
3 foo train 3 1.0 2013-01-02 1.0
============================================================
A B C D E F
3 1.0 2013-01-02 1.0 3 train foo
1 1.0 2013-01-02 1.0 3 train foo
2 1.0 2013-01-02 1.0 3 test foo
0 1.0 2013-01-02 1.0 3 test foo
============================================================
A C E
3 1.0 1.0 train
1 1.0 1.0 train
2 1.0 1.0 test
0 1.0 1.0 test
四、聚合分组
print(df2.groupby(['E'],as_index=False).mean())
print('=' * 60)
print(df2[['A','E']].groupby(['E'],as_index=False).mean())
print('=' * 60)
print(df2['A'].groupby(df2['E']).mean())
# 输出
E A C D
0 test 1.0 1.0 3
1 train 1.0 1.0 3
============================================================
E A
0 test 1.0
1 train 1.0
============================================================
E
test 1.0
train 1.0
Name: A, dtype: float64
聚合分组加排序
print(df2[['A','E']].groupby(['E'],as_index=False).sum().sort_values(by='E', ascending=False))
# 输出
E A
1 train 2.0
0 test 2.0
五、切片
1、通过[]（List）
# 用[]选取行
print(df[0:1]) # 选取一行
print('=' * 60)
print(df[0:-1]) # 选取所有行
print('=' * 60)
print(df[0:3]) # 选取前三行
print('=' * 60)
print(df['20130102':'20130104']) # 选取特定值行
# 输出
A B C D
2013-01-01 -1.584409 -1.85485 -0.284502 -0.168737
============================================================
A B C D
2013-01-01 -1.584409 -1.854850 -0.284502 -0.168737
2013-01-02 -0.182270 -1.599832 -0.242016 0.055275
2013-01-03 0.200055 0.797812 0.007122 -0.587346
2013-01-04 0.463844 0.072262 0.253212 -0.839793
2013-01-05 -0.631868 1.829623 1.135196 0.693634
============================================================
A B C D
2013-01-01 -1.584409 -1.854850 -0.284502 -0.168737
2013-01-02 -0.182270 -1.599832 -0.242016 0.055275
2013-01-03 0.200055 0.797812 0.007122 -0.587346
============================================================
A B C D
2013-01-02 -0.182270 -1.599832 -0.242016 0.055275
2013-01-03 0.200055 0.797812 0.007122 -0.587346
2013-01-04 0.463844 0.072262 0.253212 -0.839793
# 用[]选取多列
print(df[['A','B','C']])
# 输出
A B C
2013-01-01 -1.584409 -1.854850 -0.284502
2013-01-02 -0.182270 -1.599832 -0.242016
2013-01-03 0.200055 0.797812 0.007122
2013-01-04 0.463844 0.072262 0.253212
2013-01-05 -0.631868 1.829623 1.135196
2013-01-06 -1.018113 0.523377 -1.427537
# 选取单列
print(df.A)
print('=' * 60)
print(df['A']) # 效果同上
#输出
2013-01-01 -1.584409
2013-01-02 -0.182270
2013-01-03 0.200055
2013-01-04 0.463844
2013-01-05 -0.631868
2013-01-06 -1.018113
Freq: D, Name: A, dtype: float64
============================================================
2013-01-01 -1.584409
2013-01-02 -0.182270
2013-01-03 0.200055
2013-01-04 0.463844
2013-01-05 -0.631868
2013-01-06 -1.018113
Freq: D, Name: A, dtype: float64
2、通过标签（loc，at）选择
# 选取行
print(df.loc[dates[0]]) # 选取第一行
print('=' * 60)
print(df.loc[dates[0:2]]) # 选取前两行
print('=' * 60)
print(df.loc['2013-01-01':'2013-01-02']) # 同上
# 输出
A -1.584409
B -1.854850
C -0.284502
D -0.168737
Name: 2013-01-01 00:00:00, dtype: float64
============================================================
A B C D
2013-01-01 -1.584409 -1.854850 -0.284502 -0.168737
2013-01-02 -0.182270 -1.599832 -0.242016 0.055275
============================================================
A B C D
2013-01-01 -1.584409 -1.854850 -0.284502 -0.168737
2013-01-02 -0.182270 -1.599832 -0.242016 0.055275
# 选取列
print(df.loc[:,['A','B']]) # 选取A列、B列
print('=' * 60)
print(df.loc[:,'A']) # 选取A列
# 输出
A B
2013-01-01 -1.584409 -1.854850
2013-01-02 -0.182270 -1.599832
2013-01-03 0.200055 0.797812
2013-01-04 0.463844 0.072262
2013-01-05 -0.631868 1.829623
2013-01-06 -1.018113 0.523377
============================================================
2013-01-01 -1.584409
2013-01-02 -0.182270
2013-01-03 0.200055
2013-01-04 0.463844
2013-01-05 -0.631868
2013-01-06 -1.018113
Freq: D, Name: A, dtype: float64
# 同时选取行列
print(df.loc[dates[1:3],['B','C']])
print('=' * 60)
print(df.loc['2013-01-02':'2013-01-05',['B','D']])
# 输出
B C
2013-01-02 -1.599832 -0.242016
2013-01-03 0.797812 0.007122
============================================================
B D
2013-01-02 -1.599832 0.055275
2013-01-03 0.797812 -0.587346
2013-01-04 0.072262 -0.839793
2013-01-05 1.829623 0.693634
# 选取某一个特定值
print(df.loc[dates[0],'B']) # 等同于 df.at[dates[0],'B']，at不支持List
print('=' * 60)
print(df.loc['2013-01-02','D'])
# 输出
-1.8548503338248823
============================================================
0.0552752957849636
3、按索引取值
# 选取行
print(df.iloc[3])
print('=' * 60)
print(df.iloc[2:4,:])
print('=' * 60)
print(df.iloc[[2,4],:])
# 输出
A 0.463844
B 0.072262
C 0.253212
D -0.839793
Name: 2013-01-04 00:00:00, dtype: float64
============================================================
A B C D
2013-01-03 0.200055 0.797812 0.007122 -0.587346
2013-01-04 0.463844 0.072262 0.253212 -0.839793
============================================================
A B C D
2013-01-03 0.200055 0.797812 0.007122 -0.587346
2013-01-05 -0.631868 1.829623 1.135196 0.693634
# 选取列
print(df.iloc[:3])
print('=' * 60)
print(df.iloc[:,1:3])
print('=' * 60)
print(df.iloc[:,[1,3]])
# 输出
A B C D
2013-01-01 -1.584409 -1.854850 -0.284502 -0.168737
2013-01-02 -0.182270 -1.599832 -0.242016 0.055275
2013-01-03 0.200055 0.797812 0.007122 -0.587346
============================================================
B C
2013-01-01 -1.854850 -0.284502
2013-01-02 -1.599832 -0.242016
2013-01-03 0.797812 0.007122
2013-01-04 0.072262 0.253212
2013-01-05 1.829623 1.135196
2013-01-06 0.523377 -1.427537
============================================================
B D
2013-01-01 -1.854850 -0.168737
2013-01-02 -1.599832 0.055275
2013-01-03 0.797812 -0.587346
2013-01-04 0.072262 -0.839793
2013-01-05 1.829623 0.693634
2013-01-06 0.523377 -0.549401
# 同时选取多行多列
print(df.iloc[2:4,1:3])
print('=' * 60)
print(df.iloc[[2,4],[1,3]])
# 输出
B C
2013-01-03 0.797812 0.007122
2013-01-04 0.072262 0.253212
============================================================
B D
2013-01-03 0.797812 -0.587346
2013-01-05 1.829623 0.693634
# 选取某一个特定值
print(df.iloc[1,1])
print('=' * 60)
print(df.iat[1,1])
# 输出
-1.599831597530119
============================================================
-1.599831597530119
4、按布尔值取值
# 条件判断
print(df.A >0)
boolean = df.A > 0
print(df[boolean])
# 输出
2013-01-01 False
2013-01-02 False
2013-01-03 True
2013-01-04 True
2013-01-05 False
2013-01-06 False
Freq: D, Name: A, dtype: bool
A B C D
2013-01-03 0.200055 0.797812 0.007122 -0.587346
2013-01-04 0.463844 0.072262 0.253212 -0.839793
# 使用.isin()
df2 = df.copy()
df2['E'] = ['one', 'one','two','three','four','three']
print(df2[df2['E'].isin(['two','four'])])
print('=' * 60)
print(df2[~df2['E'].isin(['two','four'])])
print('=' * 60)
row_index = [x for i,x in enumerate(df2.index) if df2.iat[i,4] in ['two','four']]
print(df2.loc[row_index])
print('=' * 60)
row_index = [x for i,x in enumerate(df2.index) if df2.iat[i,4] == 'two']
print(df2.loc[row_index])
# 输出
A B C D E
2013-01-03 0.200055 0.797812 0.007122 -0.587346 two
2013-01-05 -0.631868 1.829623 1.135196 0.693634 four
============================================================
A B C D E
2013-01-01 -1.584409 -1.854850 -0.284502 -0.168737 one
2013-01-02 -0.182270 -1.599832 -0.242016 0.055275 one
2013-01-04 0.463844 0.072262 0.253212 -0.839793 three
2013-01-06 -1.018113 0.523377 -1.427537 -0.549401 three
============================================================
A B C D E
2013-01-03 0.200055 0.797812 0.007122 -0.587346 two
2013-01-05 -0.631868 1.829623 1.135196 0.693634 four
============================================================
A B C D E
2013-01-03 0.200055 0.797812 0.007122 -0.587346 two
六、缺失值处理
# Pandas默认使用np.nan来代表缺失数据。
# 创建DataFrame对象df1，以dates[0:4]为索引，在df的基础上再加一个新的列'E'（初始均为NaN）
df1 = df.reindex(index=dates[0:4], columns=list(df.columns) + ['E'])
# 将'E'列的前两个行设为1
df1.loc[dates[0]:dates[1],'E'] = 1
print(df1)
# 输出
A B C D E
2013-01-01 -1.584409 -1.854850 -0.284502 -0.168737 1.0
2013-01-02 -0.182270 -1.599832 -0.242016 0.055275 1.0
2013-01-03 0.200055 0.797812 0.007122 -0.587346 NaN
2013-01-04 0.463844 0.072262 0.253212 -0.839793 NaN
# 删除、填充缺失值
print(df1.dropna(how='any')) # 剔除df1中含NaN的行（只要任一一列出现NaN）
print('=' * 60)
print(df1.dropna(how='all')) # 剔除df1中行内值全是NaN的行
print('=' * 60)
print(df1.fillna(value=5)) # 用5填充df1里的缺失值
print('=' * 60)
print(pd.isnull(df1)) # 判断df1中的值是否为缺失数据，返回True/False
七、函数应用
1、函数
print(df.mean(0)) # 按列求平均数
print('=' * 60)
print(df.mean(1)) # 按行求平均数
# 输出
A -0.458793
B -0.038601
C -0.093088
D -0.232728
dtype: float64
============================================================
2013-01-01 -0.973125
2013-01-02 -0.492211
2013-01-03 0.104411
2013-01-04 -0.012619
2013-01-05 0.756646
2013-01-06 -0.617919
Freq: D, dtype: float64
2、apply函数
print(df)
print(df.apply(np.cumsum))
print('=' * 60)
print(df.apply(lambda x: x.max()-x.min()))
# 输出
A B C D
2013-01-01 -1.584409 -1.854850 -0.284502 -0.168737
2013-01-02 -0.182270 -1.599832 -0.242016 0.055275
2013-01-03 0.200055 0.797812 0.007122 -0.587346
2013-01-04 0.463844 0.072262 0.253212 -0.839793
2013-01-05 -0.631868 1.829623 1.135196 0.693634
2013-01-06 -1.018113 0.523377 -1.427537 -0.549401
A B C D
2013-01-01 -1.584409 -1.854850 -0.284502 -0.168737
2013-01-02 -1.766679 -3.454682 -0.526517 -0.113462
2013-01-03 -1.566624 -2.656870 -0.519395 -0.700808
2013-01-04 -1.102780 -2.584609 -0.266183 -1.540601
2013-01-05 -1.734648 -0.754986 0.869012 -0.846967
2013-01-06 -2.752761 -0.231609 -0.558525 -1.396368
============================================================
A 2.048253
B 3.684473
C 2.562733
D 1.533426
dtype: float64
八、数据集合并
1、连接
df = pd.DataFrame(np.random.randn(10,4))
pieces = [df[:3],df[3:7],df[7:]]
print(pieces)
print('=' * 60)
print(pd.concat(pieces))
# 输出
[ 0 1 2 3
0 0.639723 0.444278 0.742950 0.017251
1 -0.885244 0.728299 -0.086944 -2.956531
2 1.549633 -0.388373 0.604226 -2.166363, 0 1 2 3
3 -1.178607 -0.275841 -0.747447 1.147744
4 0.821686 0.050840 -0.567725 0.994611
5 0.908378 -1.193426 0.160251 0.701478
6 0.706225 -1.959642 -1.132494 0.672372, 0 1 2 3
7 -0.583920 -1.108532 0.118670 0.133175
8 1.801885 0.053336 -0.665703 0.628828
9 0.146570 -0.694837 -1.508931 0.583960]
============================================================
0 1 2 3
0 0.639723 0.444278 0.742950 0.017251
1 -0.885244 0.728299 -0.086944 -2.956531
2 1.549633 -0.388373 0.604226 -2.166363
3 -1.178607 -0.275841 -0.747447 1.147744
4 0.821686 0.050840 -0.567725 0.994611
5 0.908378 -1.193426 0.160251 0.701478
6 0.706225 -1.959642 -1.132494 0.672372
7 -0.583920 -1.108532 0.118670 0.133175
8 1.801885 0.053336 -0.665703 0.628828
9 0.146570 -0.694837 -1.508931 0.583960
2、追加列
df = pd.DataFrame(np.random.randn(8, 4), columns=['A','B','C','D'])
print(df)
print("=" * 60)
s = df.iloc[3]
print(df.append(s, ignore_index=True))
# 输出
A B C D
0 -0.578743 -0.150960 -0.446618 -0.988564
1 1.225723 -0.047997 -1.271592 0.035804
2 0.138365 1.788831 1.388763 -0.205168
3 1.245087 -1.409133 -1.346689 -0.012804
4 0.783840 -3.361609 1.415269 1.091229
5 -0.614037 0.302538 -1.202600 -1.220887
6 0.361368 -0.812398 -1.773582 -0.345263
7 0.636777 -0.593689 0.219574 0.239153
============================================================
A B C D
0 -0.578743 -0.150960 -0.446618 -0.988564
1 1.225723 -0.047997 -1.271592 0.035804
2 0.138365 1.788831 1.388763 -0.205168
3 1.245087 -1.409133 -1.346689 -0.012804
4 0.783840 -3.361609 1.415269 1.091229
5 -0.614037 0.302538 -1.202600 -1.220887
6 0.361368 -0.812398 -1.773582 -0.345263
7 0.636777 -0.593689 0.219574 0.239153
8 1.245087 -1.409133 -1.346689 -0.012804
九、重塑数据框
1、层次化（stack）
tuples = list(zip(*[['bar', 'bar', 'baz', 'baz', 'foo', 'foo', 'qux', 'qux'], ['one', 'two', 'one', 'two', 'one', 'two', 'one', 'two']]))
# 多重索引
index = pd.MultiIndex.from_tuples(tuples, names=['first', 'second'])
df = pd.DataFrame(np.random.randn(8, 2), index=index, columns=['A', 'B'])
print(df)
print("=" * 60)
# .stack()方法将DataFrame的列“压缩”了一级
stacked = df.stack()
print(stacked)
2、去层次化
print(stacked)
print("=" * 60)
print(stacked.unstack()) # 默认从最里边开始unstack
print("=" * 60)
print(stacked.unstack(1)) # 指定索引位置
print("=" * 60)
print(stacked.unstack(0))
# 输出
first second
bar one A 0.195572
B -0.108570
two A -1.660513
B -1.583929
baz one A 0.077588
B 2.668222
two A 0.147339
B -0.063172
foo one A 0.141091
B 1.345705
two A -0.238587
B 0.328344
qux one A -0.071885
B 0.052804
two A -0.896821
B 1.147172
dtype: float64
============================================================
A B
first second
bar one 0.195572 -0.108570
two -1.660513 -1.583929
baz one 0.077588 2.668222
two 0.147339 -0.063172
foo one 0.141091 1.345705
two -0.238587 0.328344
qux one -0.071885 0.052804
two -0.896821 1.147172
============================================================
second one two
first
bar A 0.195572 -1.660513
B -0.108570 -1.583929
baz A 0.077588 0.147339
B 2.668222 -0.063172
foo A 0.141091 -0.238587
B 1.345705 0.328344
qux A -0.071885 -0.896821
B 0.052804 1.147172
============================================================
first bar baz foo qux
second
one A 0.195572 0.077588 0.141091 -0.071885
B -0.108570 2.668222 1.345705 0.052804
two A -1.660513 0.147339 -0.238587 -0.896821
B -1.583929 -0.063172 0.328344 1.147172
3、数据透视表
# 常见参数有index，values，columns，aggfunc，都支持传递List。
df = pd.DataFrame({'A' : ['one', 'one', 'two', 'three'] * 3, 'B' : ['A', 'B', 'C'] * 4,
'C' : ['foo', 'foo', 'foo', 'bar', 'bar', 'bar'] * 2, 'D' : np.random.randn(12),
'E' : np.random.randn(12)})
print(df)
print("=" * 60)
print(pd.pivot_table(df, index=['A', 'B'], values='D', columns=['C']))
print("=" * 60)
print(pd.pivot_table(df, index=['A', 'B'], values='D', columns=['C'], aggfunc=[np.mean,len]))
# 输出
A B C D E
0 one A foo 0.163998 0.499117
1 one B foo -0.968083 -0.395698
2 two C foo -0.696637 0.263027
3 three A bar -1.765157 0.281752
4 one B bar 0.847807 0.432454
5 one C bar 1.107584 0.121205
6 two A foo 1.513330 -0.553265
7 three B foo 1.837021 0.388208
8 one C foo 0.322434 1.388784
9 one A bar -0.596184 -0.194123
10 two B bar 0.773536 -0.881527
11 three C bar 1.181760 -0.391355
============================================================
C bar foo
A B
one A -0.596184 0.163998
B 0.847807 -0.968083
C 1.107584 0.322434
three A -1.765157 NaN
B NaN 1.837021
C 1.181760 NaN
two A NaN 1.513330
B 0.773536 NaN
C NaN -0.696637
============================================================
mean len
C bar foo bar foo
A B
one A -0.596184 0.163998 1.0 1.0
B 0.847807 -0.968083 1.0 1.0
C 1.107584 0.322434 1.0 1.0
three A -1.765157 NaN 1.0 NaN
B NaN 1.837021 NaN 1.0
C 1.181760 NaN 1.0 NaN
two A NaN 1.513330 NaN 1.0
B 0.773536 NaN 1.0 NaN
C NaN -0.696637 NaN 1.0
4、列联表
print(pd.crosstab(df['A'],df['C'],margins=True,normalize=True))
# 输出
C bar foo All
A
one 0.250000 0.250000 0.50
three 0.166667 0.083333 0.25
two 0.083333 0.166667 0.25
All 0.500000 0.500000 1.00
# 也可以循环实现列联表
val_list = ['A','B','C']
for i in val_list:
print('='*100)
print(pd.crosstab(df[df['B'] == i]['B'], df[df['B'] == i]['C'],margins=True,normalize=True))
# 输出
====================================================================================================
C bar foo All
B
A 0.5 0.5 1.0
All 0.5 0.5 1.0
====================================================================================================
C bar foo All
B
B 0.5 0.5 1.0
All 0.5 0.5 1.0
====================================================================================================
C bar foo All
B
C 0.5 0.5 1.0
All 0.5 0.5 1.0
扫一扫
3万+
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
