太香了！最强的 pandas 入门教程
pip install pandas
import pandas as pd # 导入pandas库
print(pd.__version__) # 打印pandas版本信息
#> 0.23.4
myarr = np.arange(26) # 数组
mydict = dict(zip(mylist, myarr)) # 字典
# 构建方法
ser1 = pd.Series(mylist)
ser2 = pd.Series(myarr)
ser3 = pd.Series(mydict)
print(ser3.head()) # 打印前5个数据
#> a 0
b 1
c 2
d 4
e 3
dtype:int64
myarr = np.arange(26)
mydict = dict(zip(mylist, myarr))
ser = pd.Series(mydict)
# series转换为dataframe
df = ser.to_frame()
# 索引列转换为dataframe的列
df.reset_index(inplace=True)
print(df.head())
#> index 0
0 a 0
1 b 1
2 c 2
3 e 3
4 d 4
ser1 = pd.Series(list('abcedfghijklmnopqrstuvwxyz'))
# 构建series2
ser2 = pd.Series(np.arange(26))
# 方法1，axis=1表示列拼接，0表示行拼接
df = pd.concat([ser1, ser2], axis=1)
# 与方法1相比，方法2设置了列名
df = pd.DataFrame({'col1': ser1, 'col2': ser2})
print(df.head())
#> col1 col2
0 a 0
1 b 1
2 c 2
3 e 3
4 d 4
# 命名索引列名称
ser.name = 'alphabets'
# 显示前5行数据
ser.head()
#>0 a
1 b
2 c
3 e
4 d
Name: alphabets, dtype: object
ser2 = pd.Series([4, 5, 6, 7, 8])
# 返回ser1不包含ser2的布尔型series
ser3=~ser1.isin(ser2)
# 获取ser不包含ser2的元素
ser1[ser3]
#>0 1
1 2
2 3
dtype: int64
ser2 = pd.Series([4, 5, 6, 7, 8])
# 求ser1和ser2的并集
ser_u = pd.Series(np.union1d(ser1, ser2))
# 求ser1和ser2的交集
ser_i = pd.Series(np.intersect1d(ser1, ser2))
# ser_i在ser_u的补集就是ser1和ser2不相同的项
ser_u[~ser_u.isin(ser_i)]
#>0 1
1 2
2 3
5 6
6 7
7 8
dtype: int64
state = np.random.RandomState(100)
# 从均值为5标准差为25的正态分布随机抽取5个点构成series
ser = pd.Series(state.normal(10, 5, 25))
# 求ser的四分位数
np.percentile(ser, q=[0, 25, 50, 75, 100])
#> array([ 1.25117263, 7.70986507, 10.92259345, 13.36360403, 18.0949083 ])
ser = pd.Series(np.take(list('abcdefgh'), np.random.randint(8, size=30)))
# 对该series进行计数
ser.value_counts()
#>d 8
g 6
b 6
a 5
e 2
h 2
f 1
dtype: int64
# 从1~4均匀采样12个点组成series
ser = pd.Series(np.random.randint(1, 5, [12]))
# 除前两行索引对应的值不变，后几行索引对应的值为Other
ser[~ser.isin(ser.value_counts().index[:2])] = 'Other'
ser
#>0 Other
1 4
2 2
3 2
4 4
5 Other
6 Other
7 Other
8 4
9 4
10 4
11 2
dtype: object
# 离散化10个类别值,只显示前5行的数据
pd.qcut(ser, q=[0, .10, .20, .3, .4, .5, .6, .7, .8, .9, 1],
labels=['1st', '2nd', '3rd', '4th', '5th', '6th', '7th', '8th', '9th', '10th']).head()
#>
0 3rd
1 1st
2 6th
3 6th
4 9th
dtype: category
Categories (10, object): [1st < 2nd < 3rd < 4th ... 7th < 8th < 9th < 10th]
# serier类型转换numpy类型，然后重构
df = pd.DataFrame(ser.values.reshape(7,5))
print(df)
#> 0 1 2 3 4
0 1 2 1 2 5
1 1 2 4 5 2
2 1 3 3 2 8
3 8 6 4 9 6
4 2 1 1 8 5
5 3 2 8 5 6
6 1 5 5 4 6
print(ser)
# 获取值是3倍数的索引
np.argwhere(ser % 3==0)
#>0 6
1 8
2 6
3 7
4 6
5 2
6 4
dtype: int64
#>array([[0],
[2],
[4]])
index = [0, 4, 8, 14, 20]
# 获取指定索引的元素
ser.take(index)
#>0 a
4 e
8 i
14 o
20 u
dtype: object
ser2 = pd.Series(list('abcde'))
# 垂直拼接
df = pd.concat([ser1, ser2], axis=0)
# 水平拼接
df = pd.concat([ser1, ser2], axis=1)
print(df)
#> 0 1
0 0 a
1 1 b
2 2 c
3 3 d
4 4 e
ser1 = pd.Series([10, 9, 6, 5, 3, 1, 12, 8, 13])
ser2 = pd.Series([1, 3, 10, 13])
# 方法 1
[np.where(i == ser1)[0].tolist()[0] for i in ser2]
# 方法 2
[pd.Index(ser1).get_loc(i) for i in ser2]
#> [5, 4, 0, 8]
pred = pd.Series(range(10)) + np.random.random(10)
# 均方差
np.mean((truth-pred)**2)
#> 0.25508722434194103
ser = pd.Series(['how', 'to', 'kick', 'ass?'])
# 方法 1
ser.map(lambda x: x.title())
# 方法 2 ，字符串相加
ser.map(lambda x: x[0].upper() + x[1:])
# 方法 3
pd.Series([i.title() for i in ser])
#>0 How
1 To
2 Kick
3 Ass?
dtype: object
# 方法
ser.map(lambda x: len(x))
#>0 3
1 2
2 4
3 4
dtype: int64
# 求一阶导并转化为列表类型
print(ser.diff().tolist())
# 求二阶导并转化为列表类型
print(ser.diff().diff().tolist())
#>[nan, 2.0, 3.0, 4.0, 5.0, 6.0, 6.0, 8.0]
[nan, nan, 1.0, 1.0, 1.0, 1.0, 0.0, 2.0]
pd.to_datetime(ser)
#>0 2010-01-01 00:00:00
1 2011-02-02 00:00:00
2 2012-03-03 00:00:00
3 2013-04-04 00:00:00
4 2014-05-05 00:00:00
5 2015-06-06 12:20:00
dtype: datetime64[ns]
# 方法
from collections import Counter
# Counter是一个类字典类型，键是元素值，值是元素出现的次数,满足条件的元素返回True
mask = ser.map(lambda x: sum([Counter(x.lower()).get(i, 0) for i in list('aeiou')]) >= 2)
ser[mask]
#>0 Apple
1 Orange
4 Money
dtype: object
weights = pd.Series(np.linspace(1, 10, 10))
# 根据fruit对weight分组
weightsGrouped = weights.groupby(fruit)
print(weightsGrouped.indices)
# 对分组后series求每个索引的平均值
weightsGrouped.mean()
#>{'apple': array([0, 3], dtype=int64), 'banana': array([1, 2, 4, 8],
dtype=int64), 'carrot': array([5, 6, 7, 9], dtype=int64)}
#>apple 2.50
banana 4.75
carrot 7.75
dtype: float64
q = pd.Series([10, 9, 8, 7, 6, 5, 4, 3, 2, 1])
# 方法1
sum((p - q)**2)**.5
# 方法2
np.linalg.norm(p-q)
#>18.16590212458495
# 二阶导
dd = np.diff(np.sign(np.diff(ser)))
# 二阶导的最小值对应的值为最大值，返回最大值的索引
peak_locs = np.where(dd == -2)[0] + 1
peak_locs
#>array([1, 5, 7], dtype=int64)
# 方法
ser = pd.Series(list('dbc deb abed gade'))
# 统计元素的频数
freq = ser.value_counts()
print(freq)
# 求最小频数的字符
least_freq = freq.dropna().index[-1]
# 替换
''.join(ser.replace(' ', least_freq))
#>d 4
3
b 3
e 3
a 2
c 1
g 1
dtype: int64
#>'dbcgdebgabedggade'
# 求series的自相关系数，i为偏移量
autocorrelations = [ser.autocorr(i).round(2) for i in range(11)]
print(autocorrelations[1:])
# 选择最大的偏移量
print('Lag having highest correlation: ', np.argmax(np.abs(autocorrelations[1:]))+1)
#>[0.33, 0.41, 0.48, 0.01, 0.21, 0.16, -0.11, 0.05, 0.34, -0.24]
#>Lag having highest correlation: 3
import pandas as pd
series1 = pd.Series([3,4,4,4],['index1','index2','index3','index4'])
series2 = pd.Series([2,2,2,2],['index1','index2','index33','index44'])
# 加法
series_add = series1 + series2
print(series_add)
# 减法
series_minus = series1 - series2
# series_minus
# 乘法
series_multi = series1 * series2
# series_multi
# 除法
series_div = series1/series2
series_div
index1 5.0
index2 6.0
index3 NaN
index33 NaN
index4 NaN
index44 NaN
dtype: float64
#除法：
index1 1.5
index2 2.0
index3 NaN
index33 NaN
index4 NaN
index44 NaN
dtype: float64
df = pd.read_csv('https://raw.githubusercontent.com/selva86/datasets/master/Cars93_miss.csv',nrows=2,usecols=['Model','Length'])
df
#>ModelLength
0Integra177
1Legend195
df = pd.read_csv('https://raw.githubusercontent.com/selva86/datasets/master/BostonHousing.csv', chunksize=50)
df2 = pd.DataFrame()
for chunk in df:
# 获取series
df2 = df2.append(chunk.iloc[0,:])
#显示前5行
print(df2.head())
#> crim zn indus chas nox rm age \
0 0.21977 0.0 6.91 0 0.44799999999999995 5.602 62.0
1 0.0686 0.0 2.89 0 0.445 7.416 62.5
2 2.7339700000000002 0.0 19.58 0 0.871 5.597 94.9
3 0.0315 95.0 1.47 0 0.40299999999999997 6.975 15.3
4 0.19072999999999998 22.0 5.86 0 0.431 6.718 17.5
dis rad tax ptratio b lstat medv
0 6.0877 3 233 17.9 396.9 16.2 19.4
1 3.4952 2 276 18.0 396.9 6.19 33.2
2 1.5257 5 403 14.7 351.85 21.45 15.4
3 7.6534 3 402 17.0 396.9 4.56 34.9
4 7.8265 7 330 19.1 393.74 6.56 26.2
df = pd.read_csv('https://raw.githubusercontent.com/selva86/datasets/master/BostonHousing.csv',
converters={'medv': lambda x: 'High' if float(x) > 25 else 'Low'})
print(df.head())
#> b lstat medv
0 396.90 4.98 Low
1 396.90 9.14 Low
2 392.83 4.03 High
3 394.63 2.94 High
4 396.90 5.33 High
df = pd.read_csv('https://raw.githubusercontent.com/selva86/datasets/master/BostonHousing.csv', usecols=['crim', 'medv'])
# 打印前四行dataframe信息
print(df.head())
#> crim medv
0 0.00632 24.0
1 0.02731 21.6
2 0.02729 34.7
3 0.03237 33.4
4 0.06905 36.2
# 打印dataframe的行和列
print(df.shape)
# 打印dataframe每列元素的类型显示前5行
print(df.dtypes.head())
# 统计各类型的数目,方法1
print(df.get_dtype_counts())
# 统计各类型的数目,方法2
# print(df.dtypes.value_counts())
# 描述每列的统计信息，如std，四分位数等
df_stats = df.describe()
# dataframe转化数组
df_arr = df.values
# 数组转化为列表
df_list = df.values.tolist()
#>(93, 27)
Manufacturer object
Model object
Type object
Min.Price float64
Price float64
dtype: object
float64 18
object 9
dtype: int64
df = pd.read_csv('https://raw.githubusercontent.com/selva86/datasets/master/Cars93_miss.csv')
# print(df)
# 获取最大值的行和列
row, col = np.where(df.values == np.max(df.Price))
# 行和列获取最大值
print(df.iat[row[0], col[0]])
df.iloc[row[0], col[0]]
# 行索引和列名获取最大值
df.at[row[0], 'Price']
df.get_value(row[0], 'Price')
#>61.9
print(df1)
# 修改列名
print('\nchange columns :\n')
#方法1
df1.rename(columns={'weight':'stress'})
#方法2
df1.columns.values[1] = 'stress'
print(df1)
#> age weight
man1 18 50
man2 19 51
man3 20 55
change columns :
age stress
man1 18 50
man2 19 51
man3 20 55
# 若有缺失值，则为Ture
df.isnull().values.any()
#>True
# 获取每列的缺失值个数
n_missings_each_col = df.apply(lambda x: x.isnull().sum())
print(n_missings_each_col.head())
#>Manufacturer 4
Model 1
Type 3
Min.Price 7
Price 2
dtype: int64
print(df[['Min.Price','Max.Price']].head())
# 平均值替换缺失值
df_out = df[['Min.Price', 'Max.Price']] = df[['Min.Price', 'Max.Price']].apply(lambda x: x.fillna(x.mean()))
print(df_out.head())
#> Min.Price Max.Price
0 12.9 18.8
1 29.2 38.7
2 25.9 32.3
3 NaN 44.6
4 NaN NaN
#> Min.Price Max.Price
0 12.9 18.8
1 29.2 38.7
2 25.9 32.3
3 23.0 44.6
4 23.0 29.9
print(df[['Min.Price', 'Max.Price']].head())
# 全局变量
d = {'Min.Price': np.nanmean, 'Max.Price': np.nanmedian}
# 列名Min.Price的缺失值用平均值代替，Max.Price的缺失值用中值代替
df[['Min.Price', 'Max.Price']] = df[['Min.Price', 'Max.Price']].apply(lambda x, d: x.fillna(d[x.name](x)), args=(d, ))
print(df[['Min.Price', 'Max.Price']].head())
#> Min.Price Max.Price
0 12.9 18.8
1 29.2 38.7
2 25.9 32.3
3 NaN 44.6
0 12.900000 18.80
1 29.200000 38.70
2 25.900000 32.30
3 17.118605 44.60
4 17.118605 19.15
# print(df)
# 以dataframe的形式选择特定的列
type(df[['a']])
type(df.loc[:, ['a']])
print(type(df.iloc[:, [0]]))
# 以series的形式选择特定的列
type(df.a)
type(df['a'])
type(df.loc[:, 'a'])
print(type(df.iloc[:, 1]))
#><class 'pandas.core.frame.DataFrame'>
<class 'pandas.core.series.Series'>
print(df)
# 交换col1和col2
def switch_columns(df, col1=None, col2=None):
colnames = df.columns.tolist()
i1, i2 = colnames.index(col1), colnames.index(col2)
colnames[i2], colnames[i1] = colnames[i1], colnames[i2]
return df[colnames]
df1 = switch_columns(df, 'a', 'c')
print(df1)
#> a b c d e
0 0 1 2 3 4
1 5 6 7 8 9
2 10 11 12 13 14
3 15 16 17 18 19
#> c b a d e
0 2 1 0 3 4
1 7 6 5 8 9
2 12 11 10 13 14
3 17 16 15 18 19
print(df)
# 显示小数点后四位
df.apply(lambda x: '%.4f' % x, axis=1)
print(df)
#> random
0 3.539348e-04
1 3.864140e-10
2 2.973575e-02
3 1.414061e-01
#> random
0 3.539348e-04
1 3.864140e-10
2 2.973575e-02
3 1.414061e-01
# 格式化为小数点后两位的百分数
out = df.style.format({
'random': '{0:.2%}'.format,
})
out
#>random
048.54%
191.51%
290.83%
320.45%
# 每隔20行读dataframe数据
print(df.iloc[::20, :][['Manufacturer', 'Model', 'Type']])
#> Manufacturer Model Type
0 Acura Integra Small
20 Chrysler LeBaron Compact
40 Honda Prelude Sporty
60 Mercury Cougar Midsize
80 Subaru Loyale Small
print(df)
# 取'a'列前3个最大值对应的行
n = 5
df['a'].argsort()[::-1].iloc[:3]
#> a b c
0 5 5 2
1 12 7 1
2 5 2 12
3 5 14 12
4 1 13 13
#>4 1
3 3
2 2
Name: a, dtype: int64
print(df)
# dataframe每行的和
rowsums = df.apply(np.sum, axis=1)
# 选取大于100的最末两行索引
# last_two_rows = df.iloc[np.where(rowsums > 100)[0][-2:], :]
nline = np.where(rowsums > 100)[0][-2:]
nline
#> 0 1 2 3
0 19 34 15 12
1 38 35 14 26
2 39 32 18 20
3 28 27 36 38
#>array([2, 3], dtype=int64)
# 小于low_per分位的数赋值为low，大于low_per分位的数赋值为high
def cap_outliers(ser, low_perc, high_perc):
low, high = ser.quantile([low_perc, high_perc])
print(low_perc, '%ile: ', low, '|', high_perc, '%ile: ', high)
ser[ser < low] = low
ser[ser > high] = high
return(ser)
capped_ser = cap_outliers(ser, .05, .95)
#>0.05 %ile: 0.016049294076965887 | 0.95 %ile: 63.876672220183934
print(df)
# 函数
def swap_rows(df, i1, i2):
a, b = df.iloc[i1, :].copy(), df.iloc[i2, :].copy()
# 通过iloc换行
df.iloc[i1, :], df.iloc[i2, :] = b, a
return df
# 2和3行互换
print(swap_rows(df, 1, 2))
#> 0 1 2
0 0 1 2
1 3 4 5
2 6 7 8
#> 0 1 2
0 0 1 2
1 6 7 8
2 3 4 5
print(df)
# 方法 1
df.iloc[::-1, :]
# 方法 2
print(df.loc[df.index[::-1], :])
#> 0 1 2
0 0 1 2
1 3 4 5
2 6 7 8
#> 0 1 2
2 6 7 8
1 3 4 5
0 0 1 2
print(df)
# 对列'a'进行onehot编码
df_onehot = pd.concat([pd.get_dummies(df['a']), df[list('bcde')]], axis=1)
print(df_onehot)
#> a b c d e
0 0 1 2 3 4
1 5 6 7 8 9
2 10 11 12 13 14
3 15 16 17 18 19
4 20 21 22 23 24
#> 0 5 10 15 20 b c d e
0 1 0 0 0 0 1 2 3 4
1 0 1 0 0 0 6 7 8 9
2 0 0 1 0 0 11 12 13 14
3 0 0 0 1 0 16 17 18 19
4 0 0 0 0 1 21 22 23 24
print(df)
# 获取每列包含行方向上最大值的个数
count_series = df.apply(np.argmax, axis=1).value_counts()
print(count_series)
# 输出行方向最大值个数最多的列的索引
print('Column with highest row maxes: ', count_series.index[0])
#> 0 1 2
0 46 31 34
1 38 13 6
2 1 18 15
#>统计列的最大值的个数
0 2
1 1
dtype: int64
#>Column with highest row maxes: 0
# df
print(df)
# 得到四个列的相关系数
abs_corrmat = np.abs(df.corr())
print(abs_corrmat)
# 得到每个列名与其他列的最大相关系数
max_corr = abs_corrmat.apply(lambda x: sorted(x)[-2])
# 显示每列与其他列的相关系数
print('Maximum Correlation possible for each column: ', np.round(max_corr.tolist(), 2))
#> p q r s
a 59 99 1 34
b 89 60 97 40
c 43 35 14 6
d 70 59 30 53
#> p q r s
p 1.000000 0.200375 0.860051 0.744529
q 0.200375 1.000000 0.236619 0.438541
r 0.860051 0.236619 1.000000 0.341399
s 0.744529 0.438541 0.341399 1.000000
#>Maximum Correlation possible for each column: [0.86 0.44 0.86 0.74]
print(df)
# 方法1：axis=1表示行方向，
min_by_max = df.apply(lambda x: np.min(x)/np.max(x), axis=1)
# 方法2
min_by_max = np.min(df, axis=1)/np.max(df, axis=1)
min_by_max
#> 0 1 2
0 81 68 59
1 45 73 23
2 20 22 69
#>0 0.728395
1 0.315068
2 0.289855
dtype: float64
print(df)
# 行方向上取第二大的值组成series
out = df.apply(lambda x: x.sort_values().unique()[-2], axis=1)
# 构建dataframe新的列
df['penultimate'] = out
print(df)
#> 0 1 2
0 28 77 1
1 43 19 69
2 29 30 72
#> 0 1 2 penultimate
0 28 77 1 28
1 43 19 69 43
2 29 30 72 30
# 正态分布归一化
out1 = df.apply(lambda x: ((x - x.mean())/x.std()).round(2))
print('Solution Q1\n',out1)
# 线性归一化
out2 = df.apply(lambda x: ((x.max() - x)/(x.max() - x.min())).round(2))
print('Solution Q2\n', out2)
# 行与行之间的相关性
[df.iloc[i].corr(df.iloc[i+1]).round(2) for i in range(df.shape[0])[:-1]]
print(df)
# zhu
for i in range(df.shape[0]):
df.iat[i, i] = 0
df.iat[df.shape[0]-i-1, i] = 0
print(df)
#> 0 1 2 3 4
0 51 35 71 71 79
1 78 25 71 85 44
2 90 97 72 14 4
3 27 91 37 25 48
4 1 26 68 70 20
#> 0 1 2 3 4
0 0 35 71 71 0
1 78 0 71 0 44
2 90 97 0 14 4
3 27 0 37 0 48
4 0 26 68 70 0
'col2': np.random.randint(0,15,6),
'col3': np.random.randint(0, 15, 6)})
print(df)
# 按列col1分组后的平均值
df_grouped_mean = df.groupby(['col1']).mean()
print(df_grouped_mean)
# 按列col1分组后的标准差
df_grouped_std = df.groupby(['col1']).mean()
print(df_grouped_std)
#> col1 col2 col3
0 apple 2 14
1 banana 11 8
2 orange 8 10
3 apple 5 2
4 banana 6 12
5 orange 11 13
#> col2 col3
col1
apple 3.5 8.0
banana 8.5 10.0
orange 9.5 11.5
#> col2 col3
col1
apple 3.5 8.0
banana 8.5 10.0
orange 9.5 11.5
'taste': np.random.rand(6),
'price': np.random.randint(0, 15, 6)})
print(df)
# teste列按fruit分组
df_grpd = df['taste'].groupby(df.fruit)
# teste列中banana元素的信息
x=df_grpd.get_group('banana')
# 排序并找第2大的值
s = x.sort_values().iloc[-2]
print(s)
#> fruit taste price
0 apple 0.521990 7
1 banana 0.640444 0
2 orange 0.460509 9
3 apple 0.818963 4
4 banana 0.646138 7
5 orange 0.917056 12
#>0.6404436436085967
'rating': np.random.rand(6),
'price': np.random.randint(0, 15, 6)})
# 按fruit分组后，price列的平均值，并将分组置为一列
out = df.groupby('fruit', as_index=False)['price'].mean()
print(out)
#> fruit price
0 apple 4.0
1 banana 6.5
2 orange 11.0
'fruit2': np.random.choice(['apple', 'orange', 'banana'], 3)})
print(df)
# 获取两列元素相等的行
np.where(df.fruit1 == df.fruit2)
#> fruit1 fruit2
0 apple banana
1 apple apple
2 orange apple
#>(array([1], dtype=int64),)
# 创建往下偏移后的列
df['a_lag1'] = df['a'].shift(1)
# 创建往上偏移后的列
df['b_lead1'] = df['b'].shift(-1)
print(df)
#> a b c d a_lag1 b_lead1
0 29 90 43 24 NaN 36.0
1 94 36 67 66 29.0 76.0
2 81 76 44 49 94.0 97.0
3 55 97 10 74 81.0 43.0
4 32 43 62 62 55.0 NaN
# 统计元素值的个数
pd.value_counts(df.values.ravel())
#>9 3
7 3
3 3
1 3
6 2
5 2
4 2
8 1
2 1
dtype: int64
'33, Kolkata West Bengal',
'44, Chennai Tamil Nadu',
'40, Hyderabad Telengana',
'80, Bangalore Karnataka'], columns=['row'])
print(df)
# expand=True表示以分割符把字符串分成两列
df_out = df.row.str.split(',|\t', expand=True)
# 获取新的列
new_header = df_out.iloc[0]
# 重新赋值
df_out = df_out[1:]
df_out.columns = new_header
print(df_out)
0 STD, City State
1 33, Kolkata West Bengal
2 44, Chennai Tamil Nadu
3 40, Hyderabad Telengana
4 80, Bangalore Karnataka
#>0 STD City State
1 33 Kolkata West Bengal
2 44 Chennai Tamil Nadu
3 40 Hyderabad Telengana
4 80 Bangalore Karnataka
# 先通过元组方式构建多级索引
import numpy as np
outside = ['A','A','A','B','B','B']
inside =[1,2,3,1,2,3]
my_index = list(zip(outside,inside))
# my_index
# 转化为pd格式的索引
my_index = pd.MultiIndex.from_tuples(my_index)
# my_index
# 构建多级索引dataframe
df = pd.DataFrame(np.random.randn(6,2),index =my_index,columns=['fea1','fea2'])
df
#>fea1 -0.794461
fea2 0.882104
Name: 2, dtype: float64
df.loc['A'].iloc[1]['fea1']
#>-0.7944609970323794