如何从嵌套列表（是df中的列）计算频率？

Question

我想统计一个项目在嵌套列表中的出现次数。熊猫df的当前结构；每条记录都按match_id和owner_id分组，然后将第二个值（action_name，player_name）传递给名为action_seq的列表。

我可以计算每个资产中的事件总数，但是我现在希望能够计算例如玩家A曾参与活动吗？它们发生在哪些事件上的频率更高？

#sample df
pass_goal = pd.DataFrame({'match_id': [1107073,1107073,1107073,1409630,1409630], 
'possession_number': [2,2,2,40,40], 'second': [10,15,20,250,260], 
'action_name': ['pass', 'pass', 'goal','pass','goal'], 
'player_name': ['a','b','c','a','b']})

#grouping by match and possession then adding a list
posses = pass_goal.groupby(['match_id','possession_number'])[['second', 'action_name','player_name']].apply(lambda action: action.values.tolist()).reset_index(name='action_seq') 

首选输出

Player A B C
Pass   2 1 0
Goal   0 1 1

Answer 1

您可以尝试：

(pass_goal[["action_name","player_name"]]
 .pivot_table(columns="player_name", index="action_name", aggfunc=len, fill_value=0)
 .rename_axis(index="", columns="player"))

player       a  b  c

goal         0  1  1
pass         2  1  0