2017-10-21 113 views
1

我试图用np.nan值替换我的数据框中由'...'反映的缺失值。 我也想更新一些旧的值,但我的方法似乎不工作。使用Numpy和Pandas替换缺失值和更新数据帧中的旧值

这里是我的代码:

import numpy as np 
import pandas as pd 


def func(): 
    energy=pd.ExcelFile('Energy Indicators.xls').parse('Energy') 
    energy=energy.iloc[16:][['Environmental Indicators: Energy','Unnamed: 3','Unnamed: 4','Unnamed: 5']].copy() 
    energy.columns=['Country', 'Energy Supply', 'Energy Supply per Capita', '% Renewable'] 
    o="..." 
    n=np.NaN 

    # Trying to replace missing values with np.nan values 
    energy[energy['Energy Supply']==o]=n 


    energy['Energy Supply']=energy['Energy Supply']*1000000 


    # Here, I want to replace old values by new ones ==> Same problem 
    old=["Republic of Korea","United States of America","United Kingdom of " 
           +"Great Britain and Northern Ireland","China, Hong " 
           +"Kong Special Administrative Region"] 
    new=["South Korea","United States","United Kingdom","Hong Kong"] 
    for i in range(0,4): 


     energy[energy['Country']==old[i],'Country']=new[i] 


    return energy 

这里是.xls文件我的工作:https://drive.google.com/file/d/0B80lepon1RrYeDRNQVFWYVVENHM/view?usp=sharing

回答

1

我会用正则表达式做基于df.replace

energy = energy.replace(r'\s*\.+\s*', np.nan, regex=True) 

MaxU提出了一个alternative,这将工作我如果你的单元格不包含除点之外的任何特殊/空白字符。

energy = energy.replace('...', np.nan, regex=False) 
+1

我觉得应该是'能量= energy.replace( '...',np.nan,正则表达式= FALSE)' – MaxU

+0

@MaxU正则表达式默认为false,这意味着有什么事不对劲列值(可能导致空白),所以我决定去正则表达式。也会加入你的! –

+0

'energy = energy.replace('...',np.nan)'效果很好 – sali333