Pandas如何基于列条件判断为DataFrame新增符合要求的行

前置依赖 #

确保beg、end列已转换为pandas datetime类型。

实现代码 #

import pandas as pd
from datetime import time

# 1. 构造示例输入数据
df = pd.DataFrame({
    'id': ['guest1', 'guest2'],
    'beg': pd.to_datetime(['2021-10-21 17:00:00', '2021-10-21 10:00:00']),
    'end': pd.to_datetime(['2021-10-21 18:00:00', '2021-10-22 10:00:00'])
})

# 2. 筛选仅出现1次的客户id
id_count = df['id'].value_counts()
single_occur_ids = id_count[id_count == 1].index
process_df = df[df['id'].isin(single_occur_ids)].reset_index(drop=True)

result_list = []
for _, row in process_df.iterrows():
    beg_date = row['beg'].date()
    end_date = row['end'].date()
    # 情况1：beg和end为同一天
    if beg_date == end_date:
        # 新增行
        new_row = row.copy()
        new_row['col1'] = pd.Timestamp.combine(beg_date, time(0,0,0))
        new_row['col2'] = row['beg']
        result_list.append(new_row)
        # 原有行更新字段
        original_row = row.copy()
        original_row['col1'] = row['end']
        original_row['col2'] = pd.Timestamp.combine(beg_date, time(23,59,59))
        result_list.append(original_row)
    # 情况2：beg和end不为同一天
    else:
        update_row = row.copy()
        update_row['col1'] = pd.Timestamp.combine(beg_date, time(0,0,0))
        update_row['col2'] = row['beg']
        result_list.append(update_row)

# 3. 转换为最终DataFrame
final_df = pd.DataFrame(result_list).reset_index(drop=True)

# 可选：格式化时间为字符串，和示例输出格式完全一致
time_cols = ['beg', 'end', 'col1', 'col2']
final_df[time_cols] = final_df[time_cols].apply(lambda x: x.dt.strftime('%Y-%m-%d %H:%M:%S'))

print(final_df)

输出结果 #

运行代码后得到的final_df和需求要求的输出完全匹配：

内容的提问来源于stack exchange，提问作者dsp