列出熊猫集团最常见的会员？

Question

我有这样的列一个数据帧：

 id       lead_sponsor lead_sponsor_class 
    02837692 Janssen Research & Development, LLC   Industry 
    02837679    Aarhus University Hospital    Other 
    02837666 Universidad Autonoma de Ciudad Juarez    Other 
    02837653   Universidad Autonoma de Madrid    Other 
    02837640   Beirut Eye Specialist Hospital    Other 

我想找到最常见的主要发起者。我可以列出使用每个组的大小：

df.groupby(['lead_sponsor', 'lead_sponsor_class']).size() 

，给了我这样的：

lead_sponsor        lead_sponsor_class 
307 Hospital of PLA      Other     1 
3E Therapeutics Corporation    Industry    1 
3M          Industry    4 
4SC AG         Industry    8 
5 Santé         Other     1 

但我怎么找到顶级的10种最常见的群体？如果我做的：

df.groupby(['lead_sponsor', 'lead_sponsor_class']).size().sort_values(ascending=False).head(10) 

然后我得到一个错误：

AttributeError: 'Series' object has no attribute 'sort_values'

Answer 1

我认为你可以使用Series.nlargest：

print (df.groupby(['lead_sponsor', 'lead_sponsor_class']).size().nlargest(10)) 

在docs是注意：

Faster than .sort_values(ascending=False).head(n) for small n relative to the size of the Series object.

样品：

import pandas as pd 

df = pd.DataFrame({'id': {0: 2837692, 1: 2837679, 2: 2837666, 3: 2837653, 4: 2837640}, 
        'lead_sponsor': {0: 'a', 1: 'a', 2: 'a', 3: 's', 4: 's'}, 
        'lead_sponsor_class': {0: 'Industry', 1: 'Other', 2: 'Other', 3: 'Other', 4: 'Other'}}) 

print (df) 
     id lead_sponsor lead_sponsor_class 
0 2837692   a   Industry 
1 2837679   a    Other 
2 2837666   a    Other 
3 2837653   s    Other 
4 2837640   s    Other 

print (df.groupby(['lead_sponsor', 'lead_sponsor_class']).size()) 
lead_sponsor lead_sponsor_class 
a    Industry    1 
       Other     2 
s    Other     2 
dtype: int64 

print (df.groupby(['lead_sponsor', 'lead_sponsor_class']).size().sort_values(ascending=False).head(2)) 
lead_sponsor lead_sponsor_class 
s    Other     2 
a    Other     2 
dtype: int64 

print (df.groupby(['lead_sponsor', 'lead_sponsor_class']).size().nlargest(2)) 
lead_sponsor lead_sponsor_class 
a    Other     2 
s    Other     2 
dtype: int64