如何根据 dataframe 中的条件计算字符串值的出现次数？

Question

I have the following dataframe:

        Event_Type Roster_Designation
4          Assist               Male
5            Goal             Female
12         Assist             Female
13           Goal               Male
46           Goal               Male
...           ...                ...
207095       Goal             Female
207108     Assist               Male
207109       Goal               Male
207118     Assist             Female
207119       Goal             Female

What I want to know is how many Goals are scored by Females that are Assisted by Males? How many Goals scored by Females are Assisted by Females? and then vice versa (eg. Female assists -> Male goals, male assists - male goals).

Answer 1

import pandas as pd

# test data. expected result is 1 for each category
data = [['Assist', 'Male'], ['Goal', 'Female'], ['Assist', 'Female'], ['Goal', 'Male'], ['Goal', 'Male'], ['Goal', 'Female'], ['Assist', 'Male'], ['Goal', 'Male'], ['Assist', 'Female'], ['Goal', 'Female']]
df = pd.DataFrame(data, columns=['Event_Type', 'Roster_Designation'])

female_female = len(df[(df['Event_Type'] == 'Goal') & (df['Roster_Designation'] == 'Female') & (df['Roster_Designation'].shift() == 'Female') & (df['Event_Type'].shift() == 'Assist')])
female_male = len(df[(df['Event_Type'] == 'Goal') & (df['Roster_Designation'] == 'Female') & (df['Roster_Designation'].shift() == 'Male') & (df['Event_Type'].shift() == 'Assist')])
male_male = len(df[(df['Event_Type'] == 'Goal') & (df['Roster_Designation'] == 'Male') & (df['Roster_Designation'].shift() == 'Male') & (df['Event_Type'].shift() == 'Assist')])
male_female = len(df[(df['Event_Type'] == 'Goal') & (df['Roster_Designation'] == 'Male') & (df['Roster_Designation'].shift() == 'Female') & (df['Event_Type'].shift() == 'Assist')])

print(f"Female goals assisted by Females: {female_female}")
print(f"Female goals assisted by Males: {female_male}")
print(f"Male goals assisted by Males: {male_male}")
print(f"Male goals assisted by Females: {male_female}")

---

Output:
Female goals assisted by Females: 1
Female goals assisted by Males: 1
Male goals assisted by Males: 1
Male goals assisted by Females: 1

Explanation of var assignment:

We're wrapping everything in a len() so that we are assigning a number as the final value

The 4 conditions we are checking for (female -> male for example) are:

• Is current value for key Event_Type == Goal?
• Is current value for key Roster_Designation == Female?
• Is the previous value(shift) for key Roster_Designation == Male?
• Is the previous value(shift) for key Event_Type == assist?

If all above is yes, then count it.

Answer 2

How's this:

assister_index = pd.Series(
    df[df["Event_Type"] == "Goal"].index.to_series() - 1,
    index=df.index,
    dtype=pd.Int64Dtype(),
)
assister_index[~assister_index.isin(df.index)] = pd.NA

>>> assister_index
4           <NA>
5              4
12          <NA>
13            12
46          <NA>
207095      <NA>
207108      <NA>
207109    207108
207118      <NA>
207119    207118
dtype: Int64

df["Assister_Gender"] = assister_index.map(
    lambda i: df.loc[i, "Roster_Designation"], na_action="ignore"
) 
>>> df
       Event_Type Roster_Designation Assister_Gender
4          Assist               Male            <NA>
5            Goal             Female            Male
12         Assist             Female            <NA>
13           Goal               Male          Female
46           Goal               Male            <NA>
207095       Goal             Female            <NA>
207108     Assist               Male            <NA>
207109       Goal               Male            Male
207118     Assist             Female            <NA>
207119       Goal             Female          Female    

Now just count the rows matching your desired conditions, eg,

(
    (df["Event_Type"] == "Goal")
    & (df["Roster_Designation"] == "Female")
    & (df["Assister_Gender"] == "Male")
).sum()