Python Pandas 计算每年发生的事件

Question

I've got a dataframe describing events in a company and it looks like this:

employee_id    event            event_start_date    event_end_date    hire_date
1              "data change"    1.01.2018           1.01.2018         1.09.2005
2              "data change"    4.04.2018           4.04.2018         1.06.2007
2              "termination"    2.10.2020           NaT               1.06.2007
3              "hire"           23.05.2019          23.05.2019        23.05.2019
3              "leave"          23.07.2019          30.07.2019        23.05.2019
3              "termination"    3.11.2020           NaT               23.05.2019

Table is indexed by employee_id and event, and sorted by event_start_date.

So one employee has one or more events listed in the table. "Hired" event is not always in the "event" column, so I assume that information about hiring date is only available in "hire_date" column. I would like to:

1. count the number of hiring events in each year
2. count the number of termination events in each year
3. Count the number of active employees in each year

Answer 1

Build the example df:

import pandas as pd
import datetime
import numpy as np

# example df

emp = [1, 2, 2, 3, 3, 3]
event = ["data change", "data change", "termination", "hire", "leave", "termination"]
s_date = [datetime.datetime(2018, 1, 1), datetime.datetime(2018, 4, 4), datetime.datetime(2020, 10, 2),
          datetime.datetime(2019, 5, 23), datetime.datetime(2019, 7, 23), datetime.datetime(2020, 11, 3)]
e_date = [datetime.datetime(2018, 1, 1), datetime.datetime(2018, 4, 4), np.datetime64('NaT'),
          datetime.datetime(2019, 5, 23), datetime.datetime(2019, 7, 30), np.datetime64('NaT')]
h_date = [datetime.datetime(2005, 9, 1), datetime.datetime(2007, 6, 1), datetime.datetime(2017, 6, 1),
          datetime.datetime(2019, 5, 23), datetime.datetime(2019, 5, 23), datetime.datetime(2019, 5, 23)]

df = pd.DataFrame(emp, columns=['employee_id'])
df['event'] = event
df['event_start_date'] = s_date
df['event_end_date'] = e_date
df['hire_date'] = h_date

1st question

def calculate_hire_for_year():
    df['hire_year'] = pd.DatetimeIndex(df['hire_date']).year
    dict_years = {}
    ids = set(list(df['employee_id']))
    for id in ids:
        result = df[df['employee_id'] == id]
        year = list(result['hire_year'])[0]
        dict_years[year] = dict_years.get("b", 0) + 1
    return dict_years



print("Number of hiring events in each year:")
print(calculate_hire_for_year())

2nd question

def calculate_termination_per_year():
    df['year'] = pd.DatetimeIndex(df['event_start_date']).year
    result = df[df['event'] == "termination"]
    count_series = result.groupby(["event", "year"]).size()
    return count_series



print("Number of termination events in each year:")
print(calculate_termination_per_year())

3rd question

def calculate_employee_per_year():
    dict_years = {}
    df['year'] = pd.DatetimeIndex(df['event_start_date']).year
    years = set(list(df['year']))
    for year in years:
        result = df[df['year'] == year]
        count_emp = len(set(list(result['employee_id'])))
        dict_years[year] = count_emp
    return dict_years


print("Number of active employees in each year:")
print(calculate_employee_per_year())