Pandas的数据清洗——删除NaN数据

gzj137070928

删除NaN(主要针对列Series),在pandas里可以使用布尔选择或者
dropna函数删除DataFrame的某Series列里的数据，但不会影响DataFrame本身。

1. import pandas as pd
   
2. import numpy as np
   
3. val = np.arange(10, 38).reshape(7, 4)
   
4. col = list("abcd")
   
5. idx = "cake make fake sake wake lake take".split()
   
6. df = pd.DataFrame(val, columns = col, index = idx)
   
7. df["e"] = np.nan
   
8. df.at["make", "e"] = 100
   
9. df.at["wake", "e"] = 300
   
10. df.loc["jake"] = np.nan
    
11. df.at["jake", "c"] = 200
    
12. df["f"] = np.nan
    
13. print (df)
    
14. print (df.e[df.e.notnull()])
    
15. df1 = df.e.dropna()
    
16. print ('某一列删除NaN数据',df1)
    
17. df2 = df.dropna(axis=1,how='all')  # how的默认参数是any，表示只要有NaN值，则删除整行或整列
    
18. # axis=1等价于axis='column'
    
19. print ('将全部都是NaN数据的列删除',df2)
    
20. df3 = df.dropna(axis=0,how='all')  # axis的默认值是0,等价于axis='row'
    
21. print ('将全部都是NaN数据的行删除',df3)
    
22. df4 = df.copy()
    
23. df4.at['jake','c'] = np.NaN
    
24. df4.dropna(how='all',inplace=True)
    
25. print('默认删除都是NaN数据的行',df4)

复制代码