熊猫按连续值过滤？

Question

所以我试图按行中的值过滤熊猫数据框。 基本上我有一个df，其中一行包含建筑物的名称，例如。教育、K-12、办公室、教堂等。

我想根据这些值过滤一个新的数据框。 例如。我想“提取”单元格值等于“教育，K-12”的列。 我该怎么做？

我进行了广泛搜索，但大多数链式过滤似乎都基于列值。这不应该基于列值。

谢谢！

          SAN ANTONIO, TX SAN ANTONIO, TX.1 SAN ANTONIO, TX.2 SAN ANTONIO, TX.3  \
0         Commercial        Commercial        Commercial        Commercial   
1        Fossil Fuel       Fossil Fuel       Fossil Fuel       Fossil Fuel   
2    Education, K-12   Education, K-12   Education, K-12   Education, K-12   
..               ...               ...               ...               ...   
 

            SAN ANTONIO, TX.429  SAN ANTONIO, TX.430 SAN ANTONIO, TX.431  
0            Commercial          Commercial          Commercial  
1              Electric            Electric            Electric  
2         Office, Large       Office, Large       Office, Large  
..                  ...                 ...                 ...  


[745 rows x 432 columns]>

Answer 1

在测试想法后，我创建了这个

cols = df.columns[ df.iloc[2] == 'Education, K-12' ]

df[ cols ]

我只得到一行iloc[2]，所以我得到Series，我可以将Series 中的值与'Education, K-12' 进行比较 - 这为该行中的每个项目提供True/False 值，我可以使用它来过滤列。

---

最小的工作示例。

我使用io.StringIO 仅用于模拟内存中的文件，但您应该使用普通文件名。

text = '''SAN ANTONIO, TX;SAN ANTONIO, TX.1;SAN ANTONIO, TX.2;SAN ANTONIO, TX.3
Commercial;Commercial;Commercial;Commercial
Fossil Fuel;Fossil Fuel;Fossil Fuel;Fossil Fuel
Education, K-12;Education, K-12;Office, Large;Education, K-12'''

import io
import pandas as pd

df = pd.read_csv(io.StringIO(text), sep=';')

print('\n--- df ---\n')
print(df)

print('\n--- Series ---\n')
print( df.iloc[2] )

print('\n--- mask ---\n')
print( df.iloc[2] == 'Education, K-12' )

print('\n--- names ---\n')
cols = df.columns[ df.iloc[2] == 'Education, K-12' ]
print(cols)

print('\n--- columns ---\n')
print(df[ cols ])

结果：

--- df ---

   SAN ANTONIO, TX SAN ANTONIO, TX.1 SAN ANTONIO, TX.2 SAN ANTONIO, TX.3
0       Commercial        Commercial        Commercial        Commercial
1      Fossil Fuel       Fossil Fuel       Fossil Fuel       Fossil Fuel
2  Education, K-12   Education, K-12     Office, Large   Education, K-12

--- Series ---

SAN ANTONIO, TX      Education, K-12
SAN ANTONIO, TX.1    Education, K-12
SAN ANTONIO, TX.2      Office, Large
SAN ANTONIO, TX.3    Education, K-12
Name: 2, dtype: object

--- mask ---

SAN ANTONIO, TX       True
SAN ANTONIO, TX.1     True
SAN ANTONIO, TX.2    False
SAN ANTONIO, TX.3     True
Name: 2, dtype: bool

--- names ---

Index(['SAN ANTONIO, TX', 'SAN ANTONIO, TX.1', 'SAN ANTONIO, TX.3'], dtype='object')

--- columns ---

   SAN ANTONIO, TX SAN ANTONIO, TX.1 SAN ANTONIO, TX.3
0       Commercial        Commercial        Commercial
1      Fossil Fuel       Fossil Fuel       Fossil Fuel
2  Education, K-12   Education, K-12   Education, K-12

Answer 2

我以前从未见过这种用例。想不出一个优雅的方法来做，但你可以先转置数据框，然后只选择你想要的行，然后再转回来。

在下面的示例中，我将第 7 行设置为包含您要过滤的内容的行。因此，假设您要删除第 7 行中包含“c”的列。所以基本上我们需要删除“col2”。

>>> col1=['a','a','a','a','a','b','b','b']
>>> col2=['a','a','a','a','a','b','b','c']
>>> cols=['col1','col2']
>>> values=zip(col1,col2)
>>> import pandas as pd
>>> df=pd.DataFrame(data=values,columns=cols)
>>> df
  col1 col2
0    a    a
1    a    a
2    a    a
3    a    a
4    a    a
5    b    b
6    b    b
7    b    c
>>> dft=df.T
>>> dft
      0  1  2  3  4  5  6  7
col1  a  a  a  a  a  b  b  b
col2  a  a  a  a  a  b  b  c
>>> dff=dft[dft[7]!='c']
>>> dff
      0  1  2  3  4  5  6  7
col1  a  a  a  a  a  b  b  b
>>> dfo=dff.T
>>> dfo
  col1
0    a
1    a
2    a
3    a
4    a
5    b
6    b
7    b

Answer 3

我的第一个想法是转置数据帧

transposed = dt.T

在专栏中获得`教育，K-12

                            0            1                2
SAN ANTONIO, TX    Commercial  Fossil Fuel  Education, K-12
SAN ANTONIO, TX.1  Commercial  Fossil Fuel  Education, K-12
SAN ANTONIO, TX.2  Commercial  Fossil Fuel    Office, Large
SAN ANTONIO, TX.3  Commercial  Fossil Fuel  Education, K-12

然后逐行搜索

transposed[ transposed[2] == 'Education, K-12' ].index

---

最小的工作示例。

我使用io.StringIO 只是为了模拟内存中的文件，但你应该使用正常的文件名。

text = '''SAN ANTONIO, TX;SAN ANTONIO, TX.1;SAN ANTONIO, TX.2;SAN ANTONIO, TX.3
Commercial;Commercial;Commercial;Commercial
Fossil Fuel;Fossil Fuel;Fossil Fuel;Fossil Fuel
Education, K-12;Education, K-12;Office, Large;Education, K-12'''

import io
import pandas as pd

df = pd.read_csv(io.StringIO(text), sep=';')

print('\n--- df ---\n')
print(df)

transposed = df.T

print('\n--- transposed ---\n')
print(transposed)

print('\n--- names ---\n')
cols = transposed[ transposed[2] == 'Education, K-12' ].index 
print(cols)

print('\n--- columns ---\n')
print(df[ cols ])

结果

--- df ---

   SAN ANTONIO, TX SAN ANTONIO, TX.1 SAN ANTONIO, TX.2 SAN ANTONIO, TX.3
0       Commercial        Commercial        Commercial        Commercial
1      Fossil Fuel       Fossil Fuel       Fossil Fuel       Fossil Fuel
2  Education, K-12   Education, K-12     Office, Large   Education, K-12

--- transposed ---
                            0            1                2
SAN ANTONIO, TX    Commercial  Fossil Fuel  Education, K-12
SAN ANTONIO, TX.1  Commercial  Fossil Fuel  Education, K-12
SAN ANTONIO, TX.2  Commercial  Fossil Fuel    Office, Large
SAN ANTONIO, TX.3  Commercial  Fossil Fuel  Education, K-12

--- names ---

Index(['SAN ANTONIO, TX', 'SAN ANTONIO, TX.1', 'SAN ANTONIO, TX.3'], dtype='object')

--- columns ---

   SAN ANTONIO, TX SAN ANTONIO, TX.1 SAN ANTONIO, TX.3
0       Commercial        Commercial        Commercial
1      Fossil Fuel       Fossil Fuel       Fossil Fuel
2  Education, K-12   Education, K-12   Education, K-12