使用日期列值中的日期范围创建数据框

Question

给定以下数据框：

<头>

人	2018-01	2018-02	2018-03	2018-04	2018-05	2018-06	2018-07
p1		y	y		y	y	y
p2	y	y		y	y

我想返回连续“y”天的开始日期和结束日期如下：

<头>

人	开始日期	结束日期
p1	20180201	20180331
p1	20180501	20180731
p2	20180101	20180228
p2	20180401	20180531

Answer 1

假设您正在从 excel 加载数据：

import pandas as pd

# Input data prep
data = pd.read_excel('data.xlsx')
data = data.T
data.reset_index(inplace=True)

# Setting the proper header
new_header = data.iloc[0]
data = data[1:]
data.columns = new_header

# Easy to work with 1 and 0 for consecutives with cumsum
data = data.fillna(0)
data = data.replace("y", 1)


df_result = pd.DataFrame() # Store your desired table

for column in data.columns[1:]: # per person iteration
    df_temp = data[["person", column]]

    df_temp['consecutive'] = (df_temp[column].diff(1) != 0).cumsum()
    df_temp = df_temp[df_temp[column] > 0]

    df_temp = pd.DataFrame({
        'person': column,
        'start_date': df_temp.groupby('consecutive')["person"].first(),
        'end_date': df_temp.groupby('consecutive')["person"].last()
    }).reset_index(drop=True)

    df_result = df_result.append(df_temp)

# First and last day of month
df_result['start_date'] = df_result['start_date'].values.astype('datetime64[M]')
df_result['end_date'] = pd.to_datetime(df_result['end_date']) + MonthEnd(1)
print(df_result)