first commit

clean_alipay_data.py

@@ -0,0 +1,190 @@
+"""
+清理支付宝交易明细数据
+1. 仅保留指定年份的数据（默认2026年）
+2. 对于退款的条目，找到对应的支出：
+   - 如果退款金额=支出金额，两条都删除
+   - 如果退款金额<支出金额，保留差额并备注
+
+用法: python clean_alipay_data.py <输入文件> [输出文件] [--year 年份]
+示例: python clean_alipay_data.py 支付宝交易明细.csv output.csv --year 2026
+"""
+import csv
+import argparse
+from decimal import Decimal, ROUND_HALF_UP
+
+
+def parse_args():
+    parser = argparse.ArgumentParser(description="清理支付宝交易明细数据")
+    parser.add_argument("input_file", help="输入的支付宝账单CSV文件")
+    parser.add_argument("output_file", nargs="?", default=None, help="输出文件（默认为 输入文件名_cleaned.csv）")
+    parser.add_argument("--year", type=str, default="2026", help="保留的年份（默认2026）")
+    return parser.parse_args()
+
+
+def parse_amount(amount_str):
+    """解析金额字符串为Decimal"""
+    try:
+        return Decimal(amount_str.strip())
+    except:
+        return Decimal("0")
+
+
+def find_matching_expense(refund_row, expense_rows):
+    """
+    找到退款对应的支出记录
+    返回 (索引, 支出记录) 或 (None, None)
+    """
+    if len(refund_row) < 11:
+        return None, None
+
+    refund_order_no = refund_row[9].strip()  # 交易订单号
+    refund_merchant_no = refund_row[10].strip()  # 商家订单号
+
+    # 退款的交易订单号通常包含原订单号（用_分隔）
+    original_order = refund_order_no.split("_")[0] if "_" in refund_order_no else refund_order_no
+
+    for i, expense_row in enumerate(expense_rows):
+        if len(expense_row) >= 11:
+            expense_order_no = expense_row[9].strip()
+            expense_merchant_no = expense_row[10].strip()
+
+            # 匹配条件：订单号相同 或 商家订单号相同
+            if (original_order and expense_order_no == original_order) or \
+               (refund_merchant_no and expense_merchant_no == refund_merchant_no):
+                return i, expense_row
+
+    return None, None
+
+
+def main():
+    args = parse_args()
+    
+    input_file = args.input_file
+    output_file = args.output_file
+    year = args.year
+    
+    # 如果未指定输出文件，自动生成
+    if output_file is None:
+        import os
+        base_name = os.path.splitext(input_file)[0]
+        output_file = f"{base_name}_cleaned.csv"
+    
+    print(f"输入文件: {input_file}")
+    print(f"输出文件: {output_file}")
+    print(f"保留年份: {year}")
+    print()
+    
+    # 读取所有数据
+    with open(input_file, "r", encoding="utf-8") as f:
+        reader = csv.reader(f)
+        header = next(reader)
+        rows = list(reader)
+
+    print(f"原始数据行数: {len(rows)}")
+
+    # 第一步：筛选指定年份的数据
+    rows_year = []
+    for row in rows:
+        if row and row[0].startswith(year):
+            rows_year.append(row)
+
+    print(f"{year}年数据行数: {len(rows_year)}")
+
+    # 第二步：分离退款和非退款条目
+    refund_rows = []  # 退款条目
+    expense_rows = []  # 非退款条目
+
+    for row in rows_year:
+        if len(row) > 1 and row[1] == "退款":
+            refund_rows.append(row)
+        else:
+            expense_rows.append(row)
+
+    print(f"退款条目数: {len(refund_rows)}")
+    print(f"非退款条目数: {len(expense_rows)}")
+
+    # 第三步：处理退款，按订单号聚合退款金额
+    # 一个支出可能有多笔退款
+    order_refunds = {}  # 订单号 -> 退款总额
+
+    for refund_row in refund_rows:
+        if len(refund_row) >= 11:
+            refund_order_no = refund_row[9].strip()
+            refund_merchant_no = refund_row[10].strip()
+            refund_amount = parse_amount(refund_row[6])
+
+            original_order = refund_order_no.split("_")[0] if "_" in refund_order_no else refund_order_no
+
+            # 使用原订单号作为key
+            key = original_order if original_order else refund_merchant_no
+            if key:
+                if key not in order_refunds:
+                    order_refunds[key] = Decimal("0")
+                order_refunds[key] += refund_amount
+                print(f"  退款记录: {refund_row[0]} | {refund_row[2]} | {refund_amount}元")
+
+    print(f"有退款的订单数: {len(order_refunds)}")
+
+    # 第四步：处理每笔支出
+    final_rows = []
+    fully_refunded = 0
+    partially_refunded = 0
+
+    for row in expense_rows:
+        if len(row) >= 12:
+            order_no = row[9].strip()
+            merchant_no = row[10].strip()
+            expense_amount = parse_amount(row[6])
+
+            # 查找对应的退款
+            refund_amount = Decimal("0")
+            matched_key = None
+
+            for key, amount in order_refunds.items():
+                if key and (order_no == key or merchant_no == key or order_no.startswith(key)):
+                    refund_amount = amount
+                    matched_key = key
+                    break
+
+            if matched_key:
+                if refund_amount >= expense_amount:
+                    # 全额退款，删除该条目
+                    fully_refunded += 1
+                    print(f"  全额退款删除: {row[0]} | {row[2]} | {row[4][:25]}... | 原{expense_amount}元")
+                else:
+                    # 部分退款，保留差额并备注
+                    remaining = expense_amount - refund_amount
+                    remaining_str = str(remaining.quantize(Decimal("0.01"), rounding=ROUND_HALF_UP))
+                    
+                    new_row = row.copy()
+                    new_row[6] = remaining_str
+                    # 在备注列添加说明
+                    original_remark = new_row[11] if len(new_row) > 11 else ""
+                    new_row[11] = f"原金额{expense_amount}元,退款{refund_amount}元{';' + original_remark if original_remark else ''}"
+                    
+                    final_rows.append(new_row)
+                    partially_refunded += 1
+                    print(f"  部分退款: {row[0]} | {row[2]} | 原{expense_amount}元 -> {remaining_str}元")
+            else:
+                # 无退款，保留原记录
+                final_rows.append(row)
+        else:
+            final_rows.append(row)
+
+    print(f"\n处理结果:")
+    print(f"  全额退款删除: {fully_refunded} 条")
+    print(f"  部分退款调整: {partially_refunded} 条")
+    print(f"  最终保留行数: {len(final_rows)}")
+
+    # 写入清理后的数据
+    with open(output_file, "w", encoding="utf-8", newline="") as f:
+        writer = csv.writer(f)
+        writer.writerow(header)
+        writer.writerows(final_rows)
+
+    print(f"\n清理后的数据已保存到: {output_file}")
+
+
+if __name__ == "__main__":
+    main()
+