从包含 Oracle 中的 have 和 group by 子句的 Select 语句中删除行

Question

我有这个选择语句：

SELECT d.idcprefix,
         d.orgdept,
         d.idcseq,
         d.subcont, 
         d.actualcomp, 
         COUNT (*) AS "No Duplicates"
    FROM DCS_IDC, DCS_IDC z
   WHERE     D.IDCPREFIX = z.idcprefix
         AND z.orgdept = d.orgdept
         AND z.idcseq = d.idcseq
         and D.SUBCONT is not null
  HAVING COUNT (*) > 1
GROUP BY d.idcprefix,
         d.orgdept,
         d.idcseq,
         d.subcont,
         d.actualcomp
ORDER BY d.idcprefix,
         d.orgdept,
         d.idcseq,
         d.subcont,
         d.actualcomp

我想删除从该语句中返回的行，有什么想法吗？

Delete FROM 给我一个ORA-01732: data manipulation operation not legal on this view 错误，因为我相信我正在比较两个表，尽管是同一个表。

我还尝试从存在 rows 的 tablename 中删除，但删除的行比我想要的多得多。

“重复”基于 Subcont。使用 subcont 字段将记录插入到数据库/应用程序中，该字段本来是空白的，所以现在我有相同的记录，除了一个记录，例如 subcont 中的 A 和另一个为空白：

idcprefix, subcont, orgdept and idcseq
1          A        ABC         1
1                   ABC         1
2          A        BCD         1
2                   BCD         1

查询显示了所有具有 subcont 的记录，这些记录与没有 subcont 的记录重复，我需要删除具有 subcont 值的记录。

我们将不胜感激！

Answer 1

假设 (idcprefix, subcont, orgdept, idcseq) 是唯一的，那么下面的呢？

DELETE FROM DCS_IDC
WHERE (idcprefix, subcont, orgdept, idcseq) IN (
  WITH del AS (
  SELECT d.idcprefix,
     d.orgdept,
     d.idcseq,
     d.subcont, 
     d.actualcomp, 
     COUNT (*) AS "No Duplicates"
  FROM DCS_IDC, DCS_IDC z
  WHERE     D.IDCPREFIX = z.idcprefix
     AND z.orgdept = d.orgdept
     AND z.idcseq = d.idcseq
     and D.SUBCONT is not null
  HAVING COUNT (*) > 1
  GROUP BY d.idcprefix,
     d.orgdept,
     d.idcseq,
     d.subcont,
     d.actualcomp
  ORDER BY d.idcprefix,
     d.orgdept,
     d.idcseq,
     d.subcont,
     d.actualcomp
  )
 SELECT idcprefix, subcont, orgdept, idcseq FROM del
)

Answer 2

怎么样：

BEGIN
  FOR aRow IN (SELECT d.idcprefix,
                      d.orgdept,
                      d.idcseq,
                      COUNT(*) AS DUP_COUNT
                 FROM DCS_IDC d
                 HAVING COUNT(*) > 1
                 GROUP BY d.idcprefix,
                          d.orgdept,
                          d.idcseq)
  LOOP
    DELETE FROM DCS_IDC
      WHERE IDCPREFIX = aRow.IDCPREFIX AND
            ORGDEPT = aRow.ORGDEPT AND
            IDCSEQ = aRow.IDCSEQ AND
            SUBCONT IS NOT NULL;
  END LOOP;
END;

没有保证。使用风险自负。未在动物身上进行测试 - 你会是第一个！

分享和享受。