先想清楚:什么叫"重复"First: What Counts as a "Duplicate"

去重前要先定义"重复"的标准——是整行一模一样才算重复,还是只要某几列相同就算?标准不同,结果天差地别。下面三种方法对应三种标准。

Before deduping, define what "duplicate" means — identical rows, or just matching key columns? The standard changes the result completely. The three methods below map to three standards.

方法一:按整行去重Method 1: Dedupe by Whole Row

最严格:只有当一行的所有列都相同时才判定为重复。适合"同一份数据被导入了两次"这类情况。

Strictest: a row is a duplicate only when every column matches. Ideal when the same dataset was imported twice.

方法二:按指定列去重(最常用)Method 2: Dedupe by Selected Columns (Most Common)

选一个或多个"关键列"(如订单号、手机号、邮箱),只要这些列相同就视为重复,保留其中一条。这是业务去重的主力方式。

Pick one or more "key columns" (order ID, phone, email); rows matching on those are duplicates, and one is kept. This is the workhorse of business dedup.

方法三:按"近似重复"去重Method 3: Dedupe Near-Duplicates

先对关键列做去空格、统一大小写、去掉符号等标准化,再判断重复。能抓出"张三"和" 张三 "这类肉眼难辨的重复。

First normalize the key columns — trim spaces, unify case, strip symbols — then compare. Catches sneaky duplicates like "张三" vs " 张三 ".

保留哪一条?首条还是末条Keep First or Last Occurrence?

选择Choice适用场景When to use
保留首次出现Keep first后导入的是更新数据,想保留原始记录Later imports are updates; you want the original record
保留末次出现Keep last后导入的是更新数据,想保留最新记录Later imports are updates; you want the newest record

用 365Lab 去重(步骤)Dedupe with 365Lab (Steps)

  1. 在 解析页 导入 CSV。
  2. 打开 数据清洗 模块,勾选"去重"。
  3. 选择判断方式:全部列 / 指定列;并选择保留首条或末条。
  4. 必要时先用"去空格""大小写"等预处理,再用方法三。
  5. 预览确认后导出,去重操作可随时撤销。
  1. Import the CSV on the Parse page.
  2. Open the Clean module and tick "Deduplicate".
  3. Choose the basis: all columns or selected columns; keep first or last.
  4. If needed, pre-process with "trim" / "case" first, then use Method 3.
  5. Preview, export, and undo anytime.
安全提醒:去重是不可逆的"删除"操作,下手动手前建议先备份原文件;365Lab 提供最多 50 步撤销,但重新打开页面后历史会清空。
Safety note: Dedup is an irreversible delete. Back up the original first; 365Lab offers up to 50 undo steps, but history clears when you reopen the page.

清理你的重复数据 →

Clean your duplicate data →

打开数据清洗Open the Cleaner