根据各种特征（列）计算pandas的平均值

Question

Goal

I am writing a card game analysis scripts. For convenience, the data was stored in the excel sheets. So users can type in the information of each game into the excel sheets and use the python script to analyze the return of the game. 3 rivals are involved in a card game (4 person in total), and I want to analyze the overall return vs a certain player. eg. I want to know how much my dad has won when play cards with Tom.

Data

The excel sheet consists of several features like "date, start_time, end_time, duration, location, Pal1, Pal2, Pal3" and a target "Return" with positive number as gain and negative numbers as loss. The data was read using python pandas.

Problem

I did not figure out how to index a certain pal, as he/she may in one of the column "pal#". I need to calculate the mean value of return when a certain pal is involved.

Excel sheets(demo)

Code

path = 'excel.xlsx'
data_df = pd.read_excel(path)
def people_estimation(raw_data, name):
    data = raw_data
    df1 = data.pivot_table(columns=['牌友1'], values='收益', aggfunc=np.mean)
    df2 = data.pivot_table(columns=['牌友2'], values='收益', aggfunc=np.mean)
    df3 = data.pivot_table(columns=['牌友3'], values='收益', aggfunc=np.mean)
    interest = (df1[name] + df2[name] + df3[name])/3
    print("The gain with", name, "is :", interest)

Note

The code above achieve what I want. But I think there is a better way to do it. Can anyone help. Thanks in advance.

Answer 1

>>> a
   a  b  c
0  2  2  1
1  3  1  2
2  4  1  3
>>> mask = (a['a']==2) | (a['c']==2)
0     True
1     True
2    False
dtype: bool
>>> a[mask]
   a  b  c
0  2  2  1
1  3  1  2
>>> a[mask]['c']
0    1
1    2
Name: c, dtype: int64
>>> a[mask]['c'].mean()
1.5

I think in your code it's wrong that condition for a mask should by in a bracket.

data[(data['牌友1'] == 'Tom') | (data['牌友2'] == 'Tom') | (data['牌友3'] == 'Tom')]['收益'].mean()