SQL 搜索并销毁重复项

Question

我有一个包含字段的表（简化）：

id, fld1, fld2, fld3.

id 是一个数字主键字段。

存在重复项：id 不同，但 fld1、fld2 和 fld3 在 2 行或更多行中相同。当然，还有一些条目的值仅出现一次，即不重复。

在每组重复条目中，我只想保留 ID 最高的条目。我打算首先列出注定要失败的行，然后删除它们。

我的第一次尝试是这样的：

SELECT * FROM tab1 t1 WHERE EXISTS (
  SELECT COUNT(*) FROM tab1 t2
  WHERE t1.fld1 = t2.fld1 AND t1.fld2 = t2.fld2 AND t1.fld3 = t2.fld3
    AND t1.id < MAX(t2.id)
  HAVING COUNT(*) > 1
  GROUP BY t2.fld1, t2.fld2, t2.fld3)

但是（在 Oracle 中）我收到了缺少右括号错误消息。我认为这需要一种新的方法，但我的 SQL-fu 无法胜任这项任务。帮助表示赞赏！

---

编辑：

使用“真实”数据字段：

select x.leg_id, x.airline_des, x.flight_nr, x.suffix, x.flight_id_date, x.lt_flight_id_date
from fdb_leg x
join ( select max(t.leg_id) 'max_id', 
t.airline_des, t.flight_nr, t.suffix, t.flight_id_date, t.lt_flight_id_date
from fdb_leg t
group by t.airline_des, t.flight_nr, t.suffix, t.flight_id_date, t.lt_flight_id_date
having count(*) > 1) y on y.max_id > x.leg_id
and y.airline_des = x.airline_des and y.flight_nr = x.flight_nr and y.suffix = x.suffix
and y.flight_id_date = x.flight_id_date and x.lt_flight_id_date = y.lt_flight_id_date

响应为：

ORA-00923: 在预期位置未找到 FROM 关键字

Answer 1

Oracle 9i+，使用WITH：

---

要获取注定条目的列表，请使用：

WITH keepers AS (
   SELECT MAX(t.id) 'max_id',
          t.fld1, t.fld2, t.fld3
     FROM TABLE_1 t
 GROUP BY t.fld1, t.fld2, t.fld3
   HAVING COUNT(*) > 1)
SELECT x.id,
       x.fld1, x.fld2, x.fld3
  FROM TABLE_1 x
  JOIN keepers y ON y.max_id > x.id
                AND y.fld1 = x.fld1
                AND y.fld2 = x.fld2
                AND y.fld3 = x.fld3

非WITH 等效项：

---

要获取注定条目的列表，请使用：

SELECT x.id,
       x.fld1, x.fld2, x.fld3
  FROM TABLE_1 x
  JOIN (SELECT MAX(t.id) 'max_id',
               t.fld1, t.fld2, t.fld3
          FROM TABLE_1 t
      GROUP BY t.fld1, t.fld2, t.fld3
        HAVING COUNT(*) > 1) y ON y.max_id > x.id
                              AND y.fld1 = x.fld1
                              AND y.fld2 = x.fld2
                              AND y.fld3 = x.fld3

Answer 2

您可以一次性删除它们，如下所示：

SQL> create table mytable (id, fld1, fld2, fld3)
  2  as
  3  select 1, 1, 1, 1 from dual union all
  4  select 2, 1, 1, 1 from dual union all
  5  select 3, 2, 2, 2 from dual union all
  6  select 4, 2, 3, 2 from dual union all
  7  select 5, 2, 3, 2 from dual union all
  8  select 6, 2, 3, 2 from dual
  9  /

Table created.

SQL> delete mytable
  2   where id not in
  3         ( select max(id)
  4             from mytable
  5            group by fld1
  6                , fld2
  7                , fld3
  8         )
  9  /

3 rows deleted.

SQL> select * from mytable
  2  /

        ID       FLD1       FLD2       FLD3
---------- ---------- ---------- ----------
         2          1          1          1
         3          2          2          2
         6          2          3          2

3 rows selected.

问候，
抢。

Answer 3

呃，我明白了。刮掉那个。

这将识别需要删除的 ID。

Select 
       fld1
     , fld2
     , fld3
     , Max(ID)
From table_name
Group By
       fld1
     , fld2
     , fld3