【python数据分析】pandas库DataFrame基本技巧

转载请注明出处：https://blog.csdn.net/weixin_40027906/article/details/90322541

Dataframe选择行与列，切片，布尔索引

1、选择行

import pandas as pd
import numpy as np

df = pd.DataFrame(np.random.rand(12).reshape(3,4)*100)
print(df)
# 选择一行输出为Series
print('选择一行')
cl = df.loc[1]
print(cl)
print(type(cl))
# 选择多行输出为Dataframe
print('选择多行')
cl1 = df.loc[[0,1]]
print(cl1)
print(type(cl1))

-----------------------------结果---------------------------
           0          1          2          3
0  26.563019  78.730693  71.734976  69.140773
1  80.097571  11.763913  72.013929  95.420896
2   5.816948  80.052821  44.601889  18.137680
选择一行
0    80.097571
1    11.763913
2    72.013929
3    95.420896
Name: 1, dtype: float64
<class 'pandas.core.series.Series'>
选择多行
           0          1          2          3
0  26.563019  78.730693  71.734976  69.140773
1  80.097571  11.763913  72.013929  95.420896
<class 'pandas.core.frame.DataFrame'>

2、选择列

# 选择一列输出为Series
data1 = df[0]
print(data1)
print(type(data1))
# 选择多列输出为Dataframe
data2 = df[[0,1,2]]
print(data2)
print(type(data2))

-----------------------结果-----------------------
0    26.563019
1    80.097571
2     5.816948
Name: 0, dtype: float64
<class 'pandas.core.series.Series'>
           0          1          2
0  26.563019  78.730693  71.734976
1  80.097571  11.763913  72.013929
2   5.816948  80.052821  44.601889
<class 'pandas.core.frame.DataFrame'>

3、切片

import pandas as pd
import numpy as np

df1 = pd.DataFrame(np.random.rand(12).reshape(3,4)*100)
df2 = pd.DataFrame(np.random.rand(12).reshape(3,4)*100,index=list('abc'),columns=list('wxyz'))
print(df1)
print(df2)

# 按标签切片，末端包含
data1 = df1.loc[0:2]
data2 = df2.loc['a':'b']
print(data1)
print(data2)

print('-------------')
# 按照整数位切片，index的值为string时，iloc会报错
data3 = df1.iloc[0:4]
print(data3)
data4 = df2.loc['a':'b']
print(data4)

-----------------------------结果-----------------------------
           0          1          2          3
0   1.991945  72.873862  99.795669  35.488814
1  84.619386  12.016335  46.075905  90.073915
2   8.224029  79.544446  91.706230  93.593594
           w          x          y          z
a  47.279544  51.302030   1.189867  12.461972
b   1.805559  67.717741  60.762379  37.243638
c  93.600639  76.565372  14.544118  22.152201
           0          1          2          3
0   1.991945  72.873862  99.795669  35.488814
1  84.619386  12.016335  46.075905  90.073915
2   8.224029  79.544446  91.706230  93.593594
           w          x          y          z
a  47.279544  51.302030   1.189867  12.461972
b   1.805559  67.717741  60.762379  37.243638
-------------
           0          1          2          3
0   1.991945  72.873862  99.795669  35.488814
1  84.619386  12.016335  46.075905  90.073915
2   8.224029  79.544446  91.706230  93.593594
           w          x          y          z
a  47.279544  51.302030   1.189867  12.461972
b   1.805559  67.717741  60.762379  37.243638

    .loc[]与.iloc[]的区别：当index为string时，.iloc会报错

4、# Dataframe布尔索引

# Dataframe布尔索引
import pandas as pd
import numpy as np

df = pd.DataFrame(np.random.rand(12).reshape(3,4)*100)

b1 = df < 50
# 返回一个布尔型的Dataframe
print(b1)
# 返回为true的值
print(df[b1])

print('---------------')
# 指定行的布尔索引
b2 = df.loc[[0,1]] < 50
print(b2)
print(df[b2])
# 相当于
print(df[df.loc[[0,1]]<50])

print('=============')
# 指定列的布尔索引
b3 = df[[0,1,2]] > 50
print(b3)
print(df[b3])
# 相当于
print(df[df[[0,1,2]]>50])

------------------------------结果------------------------------
      0      1      2      3
0  False   True  False   True
1   True  False   True   True
2   True  False  False  False
           0          1         2          3
0        NaN  11.599345       NaN   7.074271
1  41.753027        NaN  8.381155  21.169034
2  45.363676        NaN       NaN        NaN
---------------
       0      1      2     3
0  False   True  False  True
1   True  False   True  True
           0          1         2          3
0        NaN  11.599345       NaN   7.074271
1  41.753027        NaN  8.381155  21.169034
2        NaN        NaN       NaN        NaN
           0          1         2          3
0        NaN  11.599345       NaN   7.074271
1  41.753027        NaN  8.381155  21.169034
2        NaN        NaN       NaN        NaN
=============
       0      1      2
0   True  False   True
1  False   True  False
2  False   True   True
           0          1          2   3
0  58.355777        NaN  83.006753 NaN
1        NaN  50.136687        NaN NaN
2        NaN  94.653528  59.639213 NaN
           0          1          2   3
0  58.355777        NaN  83.006753 NaN
1        NaN  50.136687        NaN NaN
2        NaN  94.653528  59.639213 NaN

5、多重索引

# 多重索引
import pandas as pd
import numpy as np

df = pd.DataFrame(np.random.rand(12).reshape(3,4)*100,
                 index=list('abc'),columns=list('wxyz'))
print(df)
print(df[['w','x']].loc[['a','b']])
print(df[df<50][['x','y']])

---------------------------结果---------------------------
           w          x          y          z
a  33.028597  36.888760  37.080400  72.748026
b  62.586162  50.126715  88.284479  91.140025
c  41.807285  94.652663  16.173990  17.088718
           w          x
a  33.028597  36.888760
b  62.586162  50.126715
          x         y
a  36.88876  37.08040
b       NaN       NaN
c       NaN  16.17399