Pandas-数据结构-DataFrame（六）：布尔型索引【用于逻辑筛选】

逃离我推掉我的手 2023-09-29 09:29 2阅读 0赞

## 一、不做索引则会对数据每个值进行判断 ##

import numpy as np
    import pandas as pd
    
    # 布尔型索引
    # 和Series原理相同
    
    df = pd.DataFrame(np.random.rand(16).reshape(4, 4) * 100,
                      index=['one', 'two', 'three', 'four'],
                      columns=['a', 'b', 'c', 'd'])
    print("df = \n", df)
    print('-' * 100)
    
    # 不做索引则会对数据每个值进行判断
    b1 = df < 20
    print("b1 = \n{0} \ntype(b1) = {1}".format(b1, type(b1)))
    print('-' * 50)
    # 索引结果保留 所有数据：True返回原数据，False返回值为NaN
    print("df[b1] = \n", df[b1])  # 也可以书写为 df[df < 20]
    print('-' * 100)

打印结果：

df = 
                    a          b          c          d
    one    87.291813  52.283368  63.663946  51.061549
    two    63.136263   3.771473  16.311522  49.151494
    three  51.227805  94.406828  68.841388  38.832457
    four   71.877744  24.139374  14.785827  40.320004
    ----------------------------------------------------------------------------------------------------
    b1 = 
               a      b      c      d
    one    False  False  False  False
    two    False   True   True  False
    three  False  False  False  False
    four   False  False   True  False 
    type(b1) = <class 'pandas.core.frame.DataFrame'>
    --------------------------------------------------
    df[b1] = 
             a         b          c   d
    one   NaN       NaN        NaN NaN
    two   NaN  3.771473  16.311522 NaN
    three NaN       NaN        NaN NaN
    four  NaN       NaN  14.785827 NaN
    ----------------------------------------------------------------------------------------------------
    
    Process finished with exit code 0

## 二、单列做判断 ##

import numpy as np
    import pandas as pd
    
    # 布尔型索引
    # 和Series原理相同
    
    df = pd.DataFrame(np.random.rand(16).reshape(4, 4) * 100,
                      index=['one', 'two', 'three', 'four'],
                      columns=['a', 'b', 'c', 'd'])
    print("df = \n", df)
    print('-' * 100)
    
    # 单列做判断
    # 索引结果保留 单列判断为True的行数据，包括其他列
    b2 = df['a'] > 50
    print("b2 = \n{0} \ntype(b2) = {1}".format(b2, type(b2)))
    print('-' * 50)
    print("df[b2] = \n", df[b2])  # 也可以书写为 df[df['a'] > 50]
    print('-' * 100)

打印结果：

df = 
                    a          b          c          d
    one    58.278853  52.279956  26.786818  58.326127
    two     2.962522  15.493893  61.898196  68.399135
    three   3.041504  58.567052  72.174175  40.690512
    four   26.269447  56.494043  24.593274   6.181143
    ----------------------------------------------------------------------------------------------------
    b2 = 
    one       True
    two      False
    three    False
    four     False
    Name: a, dtype: bool 
    type(b2) = <class 'pandas.core.series.Series'>
    --------------------------------------------------
    df[b2] = 
                  a          b          c          d
    one  58.278853  52.279956  26.786818  58.326127
    ----------------------------------------------------------------------------------------------------
    
    Process finished with exit code 0

## 三、多列做判断 ##

import numpy as np
    import pandas as pd
    
    # 布尔型索引
    # 和Series原理相同
    
    df = pd.DataFrame(np.random.rand(16).reshape(4, 4) * 100,
                      index=['one', 'two', 'three', 'four'],
                      columns=['a', 'b', 'c', 'd'])
    print("df = \n", df)
    print('-' * 100)
    
    # 多列做判断
    # 索引结果保留 所有数据：True返回原数据，False返回值为NaN
    b3 = df[['a', 'b']] > 50
    print("b3 = \n{0} \ntype(b3) = {1}".format(b3, type(b3)))
    print('-' * 50)
    print("df[b3] = \n", df[b3])  # 也可以书写为 df[df[['a','b']] > 50]
    print('-' * 100)

打印结果：

df = 
                    a          b          c          d
    one    83.682549  43.806675  18.337635   4.737157
    two    26.310925  44.744914  88.206433  78.928505
    three  17.707802  15.605249  42.536194  33.818267
    four   56.771041  76.339754   6.616507   6.201109
    ----------------------------------------------------------------------------------------------------
    b3 = 
               a      b
    one     True  False
    two    False  False
    three  False  False
    four    True   True 
    type(b3) = <class 'pandas.core.frame.DataFrame'>
    --------------------------------------------------
    df[b3] = 
                    a          b   c   d
    one    83.682549        NaN NaN NaN
    two          NaN        NaN NaN NaN
    three        NaN        NaN NaN NaN
    four   56.771041  76.339754 NaN NaN
    ----------------------------------------------------------------------------------------------------
    
    Process finished with exit code 0

## 四、多行做判断 ##

import numpy as np
    import pandas as pd
    
    # 布尔型索引
    # 和Series原理相同
    
    df = pd.DataFrame(np.random.rand(16).reshape(4, 4) * 100,
                      index=['one', 'two', 'three', 'four'],
                      columns=['a', 'b', 'c', 'd'])
    print("df = \n", df)
    print('-' * 100)
    
    # 多行做判断
    # 索引结果保留 所有数据：True返回原数据，False返回值为NaN
    b4 = df.loc[['one', 'three']] < 50
    print("b4 = \n{0} \ntype(b4) = {1}".format(b4, type(b4)))
    print('-' * 50)
    print("df[b4] = \n", df[b4])   # 也可以书写为 df[df.loc[['one','three']] < 50]
    print('-' * 100)

打印结果：

df = 
                    a          b          c          d
    one    85.557233  63.095288  82.460026  11.354019
    two    31.769011  54.641329  80.713415  59.087436
    three   9.233315  11.215880  26.345829  40.227091
    four    8.940842  76.884994  57.372274  78.014795
    ----------------------------------------------------------------------------------------------------
    b4 = 
               a      b      c     d
    one    False  False  False  True
    three   True   True   True  True 
    type(b4) = <class 'pandas.core.frame.DataFrame'>
    --------------------------------------------------
    df[b4] = 
                   a         b          c          d
    one         NaN       NaN        NaN  11.354019
    two         NaN       NaN        NaN        NaN
    three  9.233315  11.21588  26.345829  40.227091
    four        NaN       NaN        NaN        NaN
    ----------------------------------------------------------------------------------------------------
    
    Process finished with exit code 0