如何在pandas DataFrame中聚合一些数据

时间:2017-07-07 06:55:41

标签: python pandas dataframe

我有这样的数据框:

df = pd.DataFrame({'id': [115,120,200], 'category': ['a','a', 'b'], 'clust': [1, 2, 3]})

我想汇总和计算每个类别的id数量,特别是clust。例如,结果也可以是数据框,其中索引行是clust而索引列是类别,值是id的数量

1 个答案:

答案 0 :(得分:0)

IIUC,让我们使用<div id="map"> <div id="registermap"> <a href="register.html"> <img src="images/middle-earth_map1.png" alt="Registration Link" class="wobble" /> </a> </div> <div id="gollum"> <a href="cast.html"> <img src="images/gollum.png" alt="Cast and Crew Link" class="wobble" /> </a> </div> <div id="playtrailer"> <a href="trailer.html"> <img src="images/play_button.png" alt="Movie Trailer Link" class="wobble" /> </a> </div> </div>groupby

unstack

输入数据框:

import pandas as pd
df = pd.DataFrame({'id': [115,120,200], 'category': ['a', 'a', 'b'], 'clust': [1, 2, 3]})
df

分组,aggegrate和reshape:

  category  clust   id
0        a      1  115
1        a      2  120
2        b      3  200

输出:

df_out = df.groupby(['clust','category'])['id'].count().unstack()