pandas 基础操作 更新

import pandas as pd
import numpy as np

import matplotlib.pyplot as plt

创建一个Series,同时让pandas自动生成索引列

s = pd.Series([1,3,5,np.nan,6,8])

# 查看s
s

0    1.0
1    3.0
2    5.0
3    NaN
4    6.0
5    8.0
dtype: float64

创建一个DataFrame数据框

### 创建一个DataFrame ，可以传入一个numpy array 可以自己构建索引以及列标
dates = pd.date_range('2018-11-01',periods=7)
#### 比如说生成一个时间序列，以20181101 为起始位置的，7个日期组成的时间序列，数据的类型为datetime64[ns]

dates

DatetimeIndex(['2018-11-01', '2018-11-02', '2018-11-03', '2018-11-04',
               '2018-11-05', '2018-11-06', '2018-11-07'],
              dtype='datetime64[ns]', freq='D')

df = pd.DataFrame(np.random.randn(7,4),index= dates,columns=list('ABCD'))
df
# 产生随机正态分布的数据，7行4列，分别对应的index的长度以及column的长度
A B C D
2018-11-01 -0.170364 -0.237541 0.529903 0.660073
2018-11-02 -0.158446 -0.488535 0.082960 -1.913573
2018-11-03 -0.518426 0.730866 -1.033830 0.712624
2018-11-04 1.013527 0.270167 0.081805 0.178193
2018-11-05 -0.897497 -0.016279 -0.234993 0.081208
2018-11-06 -0.030580 0.545561 1.091127 -0.131579
2018-11-07 -0.313342 -0.688179 -0.417754 0.855027 
### 同时用可以使用dict的实行创建DataFrame
df2 = pd.DataFrame({"A":1,
                   "B":"20181101",
                   'C':np.array([3]*4,dtype='int32'),
                   'D':pd.Categorical(['test','train','test','train']),
                   "E":1.5},
                  )
df2
A B C D E
0 1 20181101 3 test 1.5
1 1 20181101 3 train 1.5
2 1 20181101 3 test 1.5
3 1 20181101 3 train 1.5 
df2.dtypes
### 查看数据框中的数据类型,常见的数据类型还有时间类型以及float类型

A       int64
B      object
C       int32
D    category
E     float64
dtype: object

查看数据


# 比如说看前5行
df.head()
A B C D
2018-11-01 -0.170364 -0.237541 0.529903 0.660073
2018-11-02 -0.158446 -0.488535 0.082960 -1.913573
2018-11-03 -0.518426 0.730866 -1.033830 0.712624
2018-11-04 1.013527 0.270167 0.081805 0.178193
2018-11-05 -0.897497 -0.016279 -0.234993 0.081208 
# 后4行
df.tail(4)
A B C D
2018-11-04 1.013527 0.270167 0.081805 0.178193
2018-11-05 -0.897497 -0.016279 -0.234993 0.081208
2018-11-06 -0.030580 0.545561 1.091127 -0.131579
2018-11-07 -0.313342 -0.688179 -0.417754 0.855027 
# 查看DataFrame的索引
df.index

DatetimeIndex(['2018-11-01', '2018-11-02', '2018-11-03', '2018-11-04',
               '2018-11-05', '2018-11-06', '2018-11-07'],
              dtype='datetime64[ns]', freq='D')

# 查看DataFrame的列索引
df.columns


Index(['A', 'B', 'C', 'D'], dtype='object')

# 查看DataFrame的数据,将DataFrame转化为numpy array 的数据形式
df.values

array([[-0.1703643 , -0.23754121,  0.52990284,  0.66007285],
       [-0.15844565, -0.48853537,  0.08296043, -1.91357255],
       [-0.51842554,  0.73086567, -1.03382969,  0.71262388],
       [ 1.01352712,  0.27016714,  0.08180539,  0.17819344],
       [-0.89749689, -0.01627937, -0.23499323,  0.08120819],
       [-0.03058032,  0.54556063,  1.09112723, -0.13157934],
       [-0.31334198, -0.68817881, -0.41775393,  0.85502652]])

数据的简单统计

# 可以使用describe函数对DataFrame中的数值型数据进行统计
df.describe()
A B C D
count 7.000000 7.000000 7.000000 7.000000
mean -0.153590 0.016580 0.014174 0.063139
std 0.590144 0.527860 0.680939 0.945526
min -0.897497 -0.688179 -1.033830 -1.913573
25% -0.415884 -0.363038 -0.326374 -0.025186
50% -0.170364 -0.016279 0.081805 0.178193
75% -0.094513 0.407864 0.306432 0.686348
max 1.013527 0.730866 1.091127 0.855027 
df2.describe()
### 对于其他的数据类型的数据describe函数会自动过滤掉
A C E
count 4.0 4.0 4.0
mean 1.0 3.0 1.5
std 0.0 0.0 0.0
min 1.0 3.0 1.5
25% 1.0 3.0 1.5
50% 1.0 3.0 1.5
75% 1.0 3.0 1.5
max 1.0 3.0 1.5 
### DataFrame 的转置，将列索引与行索引进行调换，行数据与列数进行调换
df.T
2018-11-01 00:00:00 2018-11-02 00:00:00 2018-11-03 00:00:00 2018-11-04 00:00:00 2018-11-05 00:00:00 2018-11-06 00:00:00 2018-11-07 00:00:00
A -0.170364 -0.158446 -0.518426 1.013527 -0.897497 -0.030580 -0.313342
B -0.237541 -0.488535 0.730866 0.270167 -0.016279 0.545561 -0.688179
C 0.529903 0.082960 -1.033830 0.081805 -0.234993 1.091127 -0.417754
D 0.660073 -1.913573 0.712624 0.178193 0.081208 -0.131579 0.855027 
df
A B C D
2018-11-01 -0.170364 -0.237541 0.529903 0.660073
2018-11-02 -0.158446 -0.488535 0.082960 -1.913573
2018-11-03 -0.518426 0.730866 -1.033830 0.712624
2018-11-04 1.013527 0.270167 0.081805 0.178193
2018-11-05 -0.897497 -0.016279 -0.234993 0.081208
2018-11-06 -0.030580 0.545561 1.091127 -0.131579
2018-11-07 -0.313342 -0.688179 -0.417754 0.855027

数据的排序

df.sort_index(ascending=False)
### 降序，按照列进行降序，通过该索引列
A B C D
2018-11-07 -0.313342 -0.688179 -0.417754 0.855027
2018-11-06 -0.030580 0.545561 1.091127 -0.131579
2018-11-05 -0.897497 -0.016279 -0.234993 0.081208
2018-11-04 1.013527 0.270167 0.081805 0.178193
2018-11-03 -0.518426 0.730866 -1.033830 0.712624
2018-11-02 -0.158446 -0.488535 0.082960 -1.913573
2018-11-01 -0.170364 -0.237541 0.529903 0.660073 

print(df.sort_values(by=['B','A']))
#  默认是升序,可以选择多指排序，先照B，后排A，如果B中的数据一样，则按照A中的大小进行排序
df.sort_values(by='B')

                   A         B         C         D
2018-11-07 -0.313342 -0.688179 -0.417754  0.855027
2018-11-02 -0.158446 -0.488535  0.082960 -1.913573
2018-11-01 -0.170364 -0.237541  0.529903  0.660073
2018-11-05 -0.897497 -0.016279 -0.234993  0.081208
2018-11-04  1.013527  0.270167  0.081805  0.178193
2018-11-06 -0.030580  0.545561  1.091127 -0.131579
2018-11-03 -0.518426  0.730866 -1.033830  0.712624
A B C D
2018-11-07 -0.313342 -0.688179 -0.417754 0.855027
2018-11-02 -0.158446 -0.488535 0.082960 -1.913573
2018-11-01 -0.170364 -0.237541 0.529903 0.660073
2018-11-05 -0.897497 -0.016279 -0.234993 0.081208
2018-11-04 1.013527 0.270167 0.081805 0.178193
2018-11-06 -0.030580 0.545561 1.091127 -0.131579
2018-11-03 -0.518426 0.730866 -1.033830 0.712624

选择数据（类似于数据库中sql语句）

df['A']
# 取出单独的一列数据，等价于df.A

2018-11-01   -0.170364
2018-11-02   -0.158446
2018-11-03   -0.518426
2018-11-04    1.013527
2018-11-05   -0.897497
2018-11-06   -0.030580
2018-11-07   -0.313342
Freq: D, Name: A, dtype: float64

# 通过[]进行行选择切片
df[0:3]
A B C D
2018-11-01 -0.170364 -0.237541 0.529903 0.660073
2018-11-02 -0.158446 -0.488535 0.082960 -1.913573
2018-11-03 -0.518426 0.730866 -1.033830 0.712624 
# 同时对于时间索引而言，可以直接使用比如
df['2018-11-01':'2018-11-04']
A B C D
2018-11-01 -0.170364 -0.237541 0.529903 0.660073
2018-11-02 -0.158446 -0.488535 0.082960 -1.913573
2018-11-03 -0.518426 0.730866 -1.033830 0.712624
2018-11-04 1.013527 0.270167 0.081805 0.178193

另外可以使用标签来选择


df.loc['2018-11-01']

A   -0.170364
B   -0.237541
C    0.529903
D    0.660073
Name: 2018-11-01 00:00:00, dtype: float64

#### 通过标签来进行多个轴上的进行选择
df.loc[:,["A","B"]] # 等价于df[["A","B"]]
A B
2018-11-01 -0.170364 -0.237541
2018-11-02 -0.158446 -0.488535
2018-11-03 -0.518426 0.730866
2018-11-04 1.013527 0.270167
2018-11-05 -0.897497 -0.016279
2018-11-06 -0.030580 0.545561
2018-11-07 -0.313342 -0.688179 
df.loc["2018-11-01":"2018-11-03",["A","B"]]
A B
2018-11-01 -0.170364 -0.237541
2018-11-02 -0.158446 -0.488535
2018-11-03 -0.518426 0.730866 
#### 获得一个标量数据
df.loc['2018-11-01','A']

-0.17036430076617162

通过位置获取数据

df.iloc[3]  # 获得第四行的数据

A    1.013527
B    0.270167
C    0.081805
D    0.178193
Name: 2018-11-04 00:00:00, dtype: float64

df.iloc[1:3,1:4]  #  与numpy中的ndarray类似
B C D
2018-11-02 -0.488535 0.08296 -1.913573
2018-11-03 0.730866 -1.03383 0.712624 
# 可以选取不连续的行或者列进行取值
df.iloc[[1,3],[1,3]]
B D
2018-11-02 -0.488535 -1.913573
2018-11-04 0.270167 0.178193 
#  对行进行切片处理
df.iloc[1:3,:]
A B C D
2018-11-02 -0.158446 -0.488535 0.08296 -1.913573
2018-11-03 -0.518426 0.730866 -1.03383 0.712624 
# 对列进行切片
df.iloc[:,1:4]
B C D
2018-11-01 -0.237541 0.529903 0.660073
2018-11-02 -0.488535 0.082960 -1.913573
2018-11-03 0.730866 -1.033830 0.712624
2018-11-04 0.270167 0.081805 0.178193
2018-11-05 -0.016279 -0.234993 0.081208
2018-11-06 0.545561 1.091127 -0.131579
2018-11-07 -0.688179 -0.417754 0.855027 
# 获取特定的值
df.iloc[1,3]

-1.9135725473596013

布尔值索引

# 使用单列的数据作为条件进行筛选
df[df.A>0]
A B C D
2018-11-04 1.013527 0.270167 0.081805 0.178193 
 #很少用到，很少使用这种大范围的条件进行筛选
df[df>0]
A B C D
2018-11-01 NaN NaN 0.529903 0.660073
2018-11-02 NaN NaN 0.082960 NaN
2018-11-03 NaN 0.730866 NaN 0.712624
2018-11-04 1.013527 0.270167 0.081805 0.178193
2018-11-05 NaN NaN NaN 0.081208
2018-11-06 NaN 0.545561 1.091127 NaN
2018-11-07 NaN NaN NaN 0.855027 
# 使用isin()方法过滤
df2.head()
A B C D E
0 1 20181101 3 test 1.5
1 1 20181101 3 train 1.5
2 1 20181101 3 test 1.5
3 1 20181101 3 train 1.5 
df2[df2['D'].isin(['test'])]
A B C D E
0 1 20181101 3 test 1.5
2 1 20181101 3 test 1.5

设定数值（类似于sql update 或者add）

  • 设定一个新的列
df['E'] = [1,2,3,4,5,6,7]

df
A B C D E
2018-11-01 -0.170364 -0.237541 0.529903 0.660073 1
2018-11-02 -0.158446 -0.488535 0.082960 -1.913573 2
2018-11-03 -0.518426 0.730866 -1.033830 0.712624 3
2018-11-04 1.013527 0.270167 0.081805 0.178193 4
2018-11-05 -0.897497 -0.016279 -0.234993 0.081208 5
2018-11-06 -0.030580 0.545561 1.091127 -0.131579 6
2018-11-07 -0.313342 -0.688179 -0.417754 0.855027 7
  • 通过标签设定新的值
df.loc['2018-11-01','E']= 10  # 第一行，E列的数据修改为10

df
A B C D E
2018-11-01 -0.170364 -0.237541 0.529903 0.660073 10
2018-11-02 -0.158446 -0.488535 0.082960 -1.913573 2
2018-11-03 -0.518426 0.730866 -1.033830 0.712624 3
2018-11-04 1.013527 0.270167 0.081805 0.178193 4
2018-11-05 -0.897497 -0.016279 -0.234993 0.081208 5
2018-11-06 -0.030580 0.545561 1.091127 -0.131579 6
2018-11-07 -0.313342 -0.688179 -0.417754 0.855027 7 
df.iloc[1,4]=5000  # 第二行第五列数据修改为5000

df
A B C D E
2018-11-01 -0.170364 -0.237541 0.529903 0.660073 10
2018-11-02 -0.158446 -0.488535 0.082960 -1.913573 5000
2018-11-03 -0.518426 0.730866 -1.033830 0.712624 3
2018-11-04 1.013527 0.270167 0.081805 0.178193 4
2018-11-05 -0.897497 -0.016279 -0.234993 0.081208 5
2018-11-06 -0.030580 0.545561 1.091127 -0.131579 6
2018-11-07 -0.313342 -0.688179 -0.417754 0.855027 7 
df3 =df.copy()
df3[df3<0]= -df3
df3  # 都变成非负数
A B C D E
2018-11-01 0.170364 0.237541 0.529903 0.660073 10
2018-11-02 0.158446 0.488535 0.082960 1.913573 5000
2018-11-03 0.518426 0.730866 1.033830 0.712624 3
2018-11-04 1.013527 0.270167 0.081805 0.178193 4
2018-11-05 0.897497 0.016279 0.234993 0.081208 5
2018-11-06 0.030580 0.545561 1.091127 0.131579 6
2018-11-07 0.313342 0.688179 0.417754 0.855027 7

缺失值处理

df
A B C D E
2018-11-01 -0.170364 -0.237541 0.529903 0.660073 10
2018-11-02 -0.158446 -0.488535 0.082960 -1.913573 5000
2018-11-03 -0.518426 0.730866 -1.033830 0.712624 3
2018-11-04 1.013527 0.270167 0.081805 0.178193 4
2018-11-05 -0.897497 -0.016279 -0.234993 0.081208 5
2018-11-06 -0.030580 0.545561 1.091127 -0.131579 6
2018-11-07 -0.313342 -0.688179 -0.417754 0.855027 7 
df['E']=[1,np.nan,2,np.nan,4,np.nan,6]

df.loc['2018-11-01':'2018-11-03','D']=np.nan

df
A B C D E
2018-11-01 -0.170364 -0.237541 0.529903 NaN 1.0
2018-11-02 -0.158446 -0.488535 0.082960 NaN NaN
2018-11-03 -0.518426 0.730866 -1.033830 NaN 2.0
2018-11-04 1.013527 0.270167 0.081805 0.178193 NaN
2018-11-05 -0.897497 -0.016279 -0.234993 0.081208 4.0
2018-11-06 -0.030580 0.545561 1.091127 -0.131579 NaN
2018-11-07 -0.313342 -0.688179 -0.417754 0.855027 6.0
  • 去掉缺失值的行
df4 = df.copy()

df4.dropna(how='any')
A B C D E
2018-11-05 -0.897497 -0.016279 -0.234993 0.081208 4.0
2018-11-07 -0.313342 -0.688179 -0.417754 0.855027 6.0 
df4.dropna(how='all')
# """DataFrame.dropna(axis=0, how='any', thresh=None, subset=None, inplace=False)""" 
# aixs 轴0或者1 index或者columns
# how 方式
# thresh 超过阈值个数的缺失值
# subset 那些字段的处理
# inplace 是否直接在原数据框中的替换
A B C D E
2018-11-01 -0.170364 -0.237541 0.529903 NaN 1.0
2018-11-02 -0.158446 -0.488535 0.082960 NaN NaN
2018-11-03 -0.518426 0.730866 -1.033830 NaN 2.0
2018-11-04 1.013527 0.270167 0.081805 0.178193 NaN
2018-11-05 -0.897497 -0.016279 -0.234993 0.081208 4.0
2018-11-06 -0.030580 0.545561 1.091127 -0.131579 NaN
2018-11-07 -0.313342 -0.688179 -0.417754 0.855027 6.0
  • 对缺失值就行填充
df4.fillna(1000)
A B C D E
2018-11-01 -0.170364 -0.237541 0.529903 1000.000000 1.0
2018-11-02 -0.158446 -0.488535 0.082960 1000.000000 1000.0
2018-11-03 -0.518426 0.730866 -1.033830 1000.000000 2.0
2018-11-04 1.013527 0.270167 0.081805 0.178193 1000.0
2018-11-05 -0.897497 -0.016279 -0.234993 0.081208 4.0
2018-11-06 -0.030580 0.545561 1.091127 -0.131579 1000.0
2018-11-07 -0.313342 -0.688179 -0.417754 0.855027 6.0
  • 对数据进行布尔值进行填充
pd.isnull(df4)
A B C D E
2018-11-01 False False False True False
2018-11-02 False False False True True
2018-11-03 False False False True False
2018-11-04 False False False False True
2018-11-05 False False False False False
2018-11-06 False False False False True
2018-11-07 False False False False False

数据操作

#统计的工作一般情况下都不包含缺失值，
df4.mean() 
#  默认是对列进行求平均，沿着行方向也就是axis=0

A   -0.153590
B    0.016580
C    0.014174
D    0.245712
E    3.250000
dtype: float64

df4.mean(axis=1)
#  沿着列方向求每行的平均

2018-11-01    0.280499
2018-11-02   -0.188007
2018-11-03    0.294653
2018-11-04    0.385923
2018-11-05    0.586488
2018-11-06    0.368632
2018-11-07    1.087150
Freq: D, dtype: float64

 # 对于拥有不同维度，需要对齐的对象进行操作。Pandas会自动的沿着指定的维度进行广播：
s = pd.Series([1,3,4,np.nan,6,7,8],index=dates)
s

2018-11-01    1.0
2018-11-02    3.0
2018-11-03    4.0
2018-11-04    NaN
2018-11-05    6.0
2018-11-06    7.0
2018-11-07    8.0
Freq: D, dtype: float64

df4.sub(s,axis='index')
A B C D E
2018-11-01 -1.170364 -1.237541 -0.470097 NaN 0.0
2018-11-02 -3.158446 -3.488535 -2.917040 NaN NaN
2018-11-03 -4.518426 -3.269134 -5.033830 NaN -2.0
2018-11-04 NaN NaN NaN NaN NaN
2018-11-05 -6.897497 -6.016279 -6.234993 -5.918792 -2.0
2018-11-06 -7.030580 -6.454439 -5.908873 -7.131579 NaN
2018-11-07 -8.313342 -8.688179 -8.417754 -7.144973 -2.0 
df4
A B C D E
2018-11-01 -0.170364 -0.237541 0.529903 NaN 1.0
2018-11-02 -0.158446 -0.488535 0.082960 NaN NaN
2018-11-03 -0.518426 0.730866 -1.033830 NaN 2.0
2018-11-04 1.013527 0.270167 0.081805 0.178193 NaN
2018-11-05 -0.897497 -0.016279 -0.234993 0.081208 4.0
2018-11-06 -0.030580 0.545561 1.091127 -0.131579 NaN
2018-11-07 -0.313342 -0.688179 -0.417754 0.855027 6.0 
df4.apply(np.cumsum)
A B C D E
2018-11-01 -0.170364 -0.237541 0.529903 NaN 1.0
2018-11-02 -0.328810 -0.726077 0.612863 NaN NaN
2018-11-03 -0.847235 0.004789 -0.420966 NaN 3.0
2018-11-04 0.166292 0.274956 -0.339161 0.178193 NaN
2018-11-05 -0.731205 0.258677 -0.574154 0.259402 7.0
2018-11-06 -0.761786 0.804237 0.516973 0.127822 NaN
2018-11-07 -1.075128 0.116059 0.099219 0.982849 13.0 
df4.apply(lambda x: x.max()-x.min())

A    1.911024
B    1.419044
C    2.124957
D    0.986606
E    5.000000
dtype: float64

统计个数与离散化

s = pd.Series(np.random.randint(0,7,size=15))
s

0     5
1     4
2     1
3     2
4     1
5     0
6     2
7     6
8     4
9     3
10    1
11    1
12    1
13    3
14    2
dtype: int32

s.value_counts()
# 统计元素的个数，并按照元素统计量进行排序，未出现的元素不会显示出来

1    5
2    3
4    2
3    2
6    1
5    1
0    1
dtype: int64

s.reindex(range(0,7))
# 按照固定的顺序输出元素的个数统计

0    5
1    4
2    1
3    2
4    1
5    0
6    2
dtype: int32

s.mode()
#  众数 

0    1
dtype: int32
  • 离散化
# 连续值转化为离散值，可以使用cut函数进行操作（bins based on vlaues） qcut （bins based on sample
# quantiles） 函数
arr = np.random.randint(0,20,size=15)  # 正态分布
arr

array([ 5, 18, 13, 16, 16,  1, 15, 11,  0, 17, 16, 18, 15, 12, 13])

factor = pd.cut(arr,3)
factor

[(-0.018, 6.0], (12.0, 18.0], (12.0, 18.0], (12.0, 18.0], (12.0, 18.0], ..., (12.0, 18.0], (12.0, 18.0], (12.0, 18.0], (6.0, 12.0], (12.0, 18.0]]
Length: 15
Categories (3, interval[float64]): [(-0.018, 6.0] < (6.0, 12.0] < (12.0, 18.0]]

pd.value_counts(factor)

(12.0, 18.0]     10
(-0.018, 6.0]     3
(6.0, 12.0]       2
dtype: int64

factor1 = pd.cut(arr,[-1,5,10,15,20])

pd.value_counts(factor1)

(15, 20]    6
(10, 15]    6
(-1, 5]     3
(5, 10]     0
dtype: int64

factor2 = pd.qcut(arr,[0,0.25,0.5,0.75,1])

pd.value_counts(factor2)

(11.5, 15.0]      5
(-0.001, 11.5]    4
(16.0, 18.0]      3
(15.0, 16.0]      3
dtype: int64

pandas 处理字符串（单独一个大的章节，这人不做详述）

数据合并

  • concat
  • merge（类似于sql数据库中的join）
  • append

首先看concat合并数据框

df = pd.DataFrame(np.random.randn(10,4))  #  10行列的标准正态分布数据框
df
0 1 2 3
0 0.949746 -0.050767 1.478622 -0.239901
1 -0.297120 -0.562589 0.371837 1.180715
2 0.953856 0.492295 0.821156 -0.323328
3 0.016153 1.554225 -1.166304 -0.904040
4 0.204763 -0.951291 -1.317620 0.672900
5 2.241006 -0.925746 -1.961408 0.853367
6 2.217133 -0.430812 0.518926 1.741445
7 -0.571104 -0.437305 -0.902241 0.786231
8 -2.511387 0.523760 1.811622 -0.777296
9 0.252690 0.901952 0.619614 -0.006631 
d1,d2,d3  = df[:3],df[3:7],df[7:]
d1,d2,d3

(          0         1         2         3
 0  0.949746 -0.050767  1.478622 -0.239901
 1 -0.297120 -0.562589  0.371837  1.180715
 2  0.953856  0.492295  0.821156 -0.323328,
           0         1         2         3
 3  0.016153  1.554225 -1.166304 -0.904040
 4  0.204763 -0.951291 -1.317620  0.672900
 5  2.241006 -0.925746 -1.961408  0.853367
 6  2.217133 -0.430812  0.518926  1.741445,
           0         1         2         3
 7 -0.571104 -0.437305 -0.902241  0.786231
 8 -2.511387  0.523760  1.811622 -0.777296
 9  0.252690  0.901952  0.619614 -0.006631)

pd.concat([d1,d2,d3])
#合并三个数据框，数据结构相同，通常合并相同结构的数据，数据框中的字段一致，类似于数据添加新的数据来源
0 1 2 3
0 0.949746 -0.050767 1.478622 -0.239901
1 -0.297120 -0.562589 0.371837 1.180715
2 0.953856 0.492295 0.821156 -0.323328
3 0.016153 1.554225 -1.166304 -0.904040
4 0.204763 -0.951291 -1.317620 0.672900
5 2.241006 -0.925746 -1.961408 0.853367
6 2.217133 -0.430812 0.518926 1.741445
7 -0.571104 -0.437305 -0.902241 0.786231
8 -2.511387 0.523760 1.811622 -0.777296
9 0.252690 0.901952 0.619614 -0.006631

merge方式合并（数据库中的join）

left = pd.DataFrame({'key':['foo','foo'],"lval":[1,2]})
right = pd.DataFrame({'key':['foo','foo'],'rval':[4,5]})

left
key lval
0 foo 1
1 foo 2 
right
key rval
0 foo 4
1 foo 5 
pd.merge(left,right,on='key')
key lval rval
0 foo 1 4
1 foo 1 5
2 foo 2 4
3 foo 2 5 
left = pd.DataFrame({'key':['foo','bar'],"lval":[1,2]})
right = pd.DataFrame({'key':['foo','bar'],'rval':[4,5]})
pd.merge(left,right,on='key')
key lval rval
0 foo 1 4
1 bar 2 5 
left
key lval
0 foo 1
1 bar 2 
right
key rval
0 foo 4
1 bar 5

Append方式合并数据

#  与concat 类似，常用的方法可以参考一下日子
df = pd.DataFrame(np.random.randn(8,4),columns=['A','B','C','D'])
df
A B C D
0 1.825997 -0.331086 -0.067143 0.747226
1 -0.027497 0.861639 0.928621 -2.549617
2 -0.546645 -0.072253 -0.788483 0.484140
3 -0.472240 -1.776993 -1.647407 0.170596
4 -0.099453 0.380143 -0.890510 1.233741
5 0.351915 0.137522 -1.165938 1.128146
6 0.558442 -1.047060 -0.598197 -1.979876
7 0.067321 -1.037666 -1.140675 -0.098562 
## 
d1 = df.iloc[3]
df.append(d1,ignore_index= True)
A B C D
0 1.825997 -0.331086 -0.067143 0.747226
1 -0.027497 0.861639 0.928621 -2.549617
2 -0.546645 -0.072253 -0.788483 0.484140
3 -0.472240 -1.776993 -1.647407 0.170596
4 -0.099453 0.380143 -0.890510 1.233741
5 0.351915 0.137522 -1.165938 1.128146
6 0.558442 -1.047060 -0.598197 -1.979876
7 0.067321 -1.037666 -1.140675 -0.098562
8 -0.472240 -1.776993 -1.647407 0.170596

分组操作Groupby操作

df = pd.DataFrame({"A":['foo','bar','foo','bar'],
                  "B":['one','one','two','three'],
                  "C":np.random.randn(4),
                  "D":np.random.randn(4)})
df
A B C D
0 foo one 0.938910 0.505163
1 bar one 0.660543 0.353860
2 foo two 0.520309 1.157462
3 bar three -1.054927 0.290693 
df.groupby('A').sum()
C D
A
bar -0.394384 0.644553
foo 1.459219 1.662625 
df.groupby('A').size()

A
bar    2
foo    2
dtype: int64

df.groupby(['A',"B"]).sum()
C D
A B
bar one 0.660543 0.353860
three -1.054927 0.290693
foo one 0.938910 0.505163
two 0.520309 1.157462 
df.groupby(['A',"B"]).size()

A    B    
bar  one      1
     three    1
foo  one      1
     two      1
dtype: int64

reshape操作

tuples = list(zip(*[['bar','bar','baz','baz','foo','foo','qux','qux'],
                   ['one','two','one','two','one','two','one','two']]))

index = pd.MultiIndex.from_tuples(tuples,names=['first','second'])
df = pd.DataFrame(np.random.randn(8,2),index= index,columns=['A','B'])
df2 =  df[:4]

df2
A B
first second
bar one 0.510758 0.641370
two 0.481230 -0.470894
baz one -0.076294 0.121247
two 0.378507 -1.358932 
df
A B
first second
bar one 0.510758 0.641370
two 0.481230 -0.470894
baz one -0.076294 0.121247
two 0.378507 -1.358932
foo one -0.873012 0.531595
two 0.266968 -0.393124
qux one 0.981866 1.205994
two 0.265772 0.132489

stack 与unstack 方法

df2_stacked = df2.stack()  
#  将column也作为index

df2_stacked

first  second   
bar    one     A    0.510758
               B    0.641370
       two     A    0.481230
               B   -0.470894
baz    one     A   -0.076294
               B    0.121247
       two     A    0.378507
               B   -1.358932
dtype: float64

df2_stacked.unstack()  #  回复到原来的状态
A B
first second
bar one 0.510758 0.641370
two 0.481230 -0.470894
baz one -0.076294 0.121247
two 0.378507 -1.358932 
df2_stacked

first  second   
bar    one     A    0.510758
               B    0.641370
       two     A    0.481230
               B   -0.470894
baz    one     A   -0.076294
               B    0.121247
       two     A    0.378507
               B   -1.358932
dtype: float64

df2_stacked.unstack(1)
second one two
first
bar A 0.510758 0.481230
B 0.641370 -0.470894
baz A -0.076294 0.378507
B 0.121247 -1.358932 
df2_stacked.unstack(0)
first bar baz
second
one A 0.510758 -0.076294
B 0.641370 0.121247
two A 0.481230 0.378507
B -0.470894 -1.358932

pivot_table 透视表

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)})

df
A B C D E
0 one A foo 0.006247 -0.894827
1 one B foo 1.653974 -0.340107
2 two C foo -1.627485 -1.011403
3 three A bar -0.716002 1.533422
4 one B bar 0.422688 -0.807675
5 one C bar 0.264818 0.249770
6 two A foo 0.643288 -1.166616
7 three B foo 0.348041 -0.659099
8 one C foo 1.593486 -1.098731
9 one A bar -0.389344 0.919528
10 two B bar -1.407450 1.269716
11 three C bar -0.172672 0.883970 
pd.pivot_table(df,values='D',index=['A','B'],columns=['C'],aggfunc=np.mean)
C bar foo
A B
one A -0.389344 0.006247
B 0.422688 1.653974
C 0.264818 1.593486
three A -0.716002 NaN
B NaN 0.348041
C -0.172672 NaN
two A NaN 0.643288
B -1.407450 NaN
C NaN -1.627485 
pd.pivot_table(df,values='D',index=['A','B'],columns=['C'],aggfunc=np.sum)
C bar foo
A B
one A -0.389344 0.006247
B 0.422688 1.653974
C 0.264818 1.593486
three A -0.716002 NaN
B NaN 0.348041
C -0.172672 NaN
two A NaN 0.643288
B -1.407450 NaN
C NaN -1.627485 
pd.pivot_table(df,values='D',index=['A','B'],columns=['C'],aggfunc=np.mean,fill_value=0)
C bar foo
A B
one A -0.389344 0.006247
B 0.422688 1.653974
C 0.264818 1.593486
three A -0.716002 0.000000
B 0.000000 0.348041
C -0.172672 0.000000
two A 0.000000 0.643288
B -1.407450 0.000000
C 0.000000 -1.627485 
df1 = pd.pivot_table(df,values='D',index=['A','B'],columns=['C'],aggfunc=np.mean,fill_value=0)

df1.index

MultiIndex(levels=[['one', 'three', 'two'], ['A', 'B', 'C']],
           labels=[[0, 0, 0, 1, 1, 1, 2, 2, 2], [0, 1, 2, 0, 1, 2, 0, 1, 2]],
           names=['A', 'B'])

df1.stack()

A      B  C  
one    A  bar   -0.389344
          foo    0.006247
       B  bar    0.422688
          foo    1.653974
       C  bar    0.264818
          foo    1.593486
three  A  bar   -0.716002
          foo    0.000000
       B  bar    0.000000
          foo    0.348041
       C  bar   -0.172672
          foo    0.000000
two    A  bar    0.000000
          foo    0.643288
       B  bar   -1.407450
          foo    0.000000
       C  bar    0.000000
          foo   -1.627485
dtype: float64

df1.unstack()
C bar foo
B A B C A B C
A
one -0.389344 0.422688 0.264818 0.006247 1.653974 1.593486
three -0.716002 0.000000 -0.172672 0.000000 0.348041 0.000000
two 0.000000 -1.407450 0.000000 0.643288 0.000000 -1.627485 
df1.unstack(1)
C bar foo
B A B C A B C
A
one -0.389344 0.422688 0.264818 0.006247 1.653974 1.593486
three -0.716002 0.000000 -0.172672 0.000000 0.348041 0.000000
two 0.000000 -1.407450 0.000000 0.643288 0.000000 -1.627485 
df1.unstack(0)
C bar foo
A one three two one three two
B
A -0.389344 -0.716002 0.00000 0.006247 0.000000 0.643288
B 0.422688 0.000000 -1.40745 1.653974 0.348041 0.000000
C 0.264818 -0.172672 0.00000 1.593486 0.000000 -1.627485

至此，pandas的基础的使用介绍也就结束了，后续会有专题性质的分析，包括(字符串处理，apply的使用，数据合并，透视表，时间序列的分析）

【推广】 免费学中医，健康全家人
原文地址：https://www.cnblogs.com/onemorepoint/p/10093098.html