如何用Python Pandas循环结合条件筛选匹配日期的银行审批数据

现有一份银行审批数据集，包含EOM Date（数据上传日期）、Approve Date（实际审批日期）及Facility Id字段。原本可以用numpy的np.where方法筛选出EOM Date与Approve Date匹配的记录，现在需要改用循环方式实现该筛选需求。

原始数据集 #

import pandas as pd

df = pd.DataFrame(data=[["1/08/23","1/08/23","yy1"], 
                    ["1/08/23","1/08/23","yy2"], 
                    ["1/08/23","1/07/23","yy3"],                                            
                    ["1/07/23","1/07/23","yy4"], 
                    ["1/07/23", "1/06/23", "yy5"], 
                    ["1/07/23","1/07/23","yy6"], 
                    ["1/06/23","1/06/23","yy7"], 
                    ["1/06/23","1/05/23","yy8"], 
                    ["1/06/23","1/04/23","yy9"]], 
                columns= ["EOM Date","Approve Date","Facility Id"])

期望输出 #

输出代码 #

filtered_df = pd.DataFrame(data=[["1/08/23","1/08/23","yy1"],
                    ["1/08/23","1/08/23","yy2"],
                    ["1/07/23","1/07/23","yy4"], 
                    ["1/07/23", "1/07/23", "yy6"],
                    ["1/06/23","1/06/23","yy7"]], 
                columns= ["EOM Date","Approve Date","Facility Id"])

输出表格 #

循环实现方案 #

import pandas as pd

# 原始数据集
df = pd.DataFrame(data=[["1/08/23","1/08/23","yy1"], 
                    ["1/08/23","1/08/23","yy2"], 
                    ["1/08/23","1/07/23","yy3"],                                            
                    ["1/07/23","1/07/23","yy4"], 
                    ["1/07/23", "1/06/23", "yy5"], 
                    ["1/07/23","1/07/23","yy6"], 
                    ["1/06/23","1/06/23","yy7"], 
                    ["1/06/23","1/05/23","yy8"], 
                    ["1/06/23","1/04/23","yy9"]], 
                columns= ["EOM Date","Approve Date","Facility Id"])

# 初始化空列表存储符合条件的行
filtered_rows = []

# 遍历DataFrame的每一行
for index, row in df.iterrows():
    # 判断EOM Date和Approve Date是否相等
    if row["EOM Date"] == row["Approve Date"]:
        filtered_rows.append(row.tolist())

# 生成筛选后的DataFrame
filtered_df = pd.DataFrame(filtered_rows, columns=df.columns)

print(filtered_df)

说明 #

• 通过df.iterrows()遍历原始数据集的每一行，获取行索引和行数据
• 逐行判断EOM Date与Approve Date是否匹配，匹配则将该行转为列表加入filtered_rows
• 最后用筛选后的行列表创建新的DataFrame，保持原列名不变

内容的提问来源于stack exchange，提问作者svenvuko