Question

是否有一种简单的方法可以在联合中跳过NULL行？我在DB2上。

示例：

SELECT 'FIRST', sum(value) FROM  SYSIBM.SYSDUMMY1
UNION
SELECT 'SECOND', sum(value) FROM  SYSIBM.SYSDUMMY1
UNION
SELECT 'THIRD', sum(value) FROM  SYSIBM.SYSDUMMY1

结果：

1---------|2-----
    FIRST   1
    SECOND  NULL/0
    THIRD   3

因此，如果其中一个联合有0或NULL，则应跳过它：

估算的示例结果：

1---------|2-----
    FIRST   1
    THIRD   3

Answer 1

使用子查询。我还建议union all：

SELECT col, val
FROM (SELECT 'FIRST' as col, sum(value) as val FROM SYSIBM.SYSDUMMY1
      UNION ALL
      SELECT 'SECOND' as col, sum(value) as val FROM SYSIBM.SYSDUMMY1
      UNION ALL
      SELECT 'THIRD' as col, sum(value) as val FROM SYSIBM.SYSDUMMY1
     ) t
WHERE val IS NOT NULL AND val > 0;

注意：

UNION会产生删除重复项的开销。使用UNION ALL可以避免这种开销。
比较val IS NOT NULL是多余的。但它明确表达了你想要的条件。

Answer 2

是的，您可以使用IS NOT NULL过滤器检查HAVING来执行此操作。

SELECT 'FIRST', sum(value)
    FROM  SYSIBM.SYSDUMMY1 GROUP BY 'FIRST' HAVING sum(value) IS NOT NULL

UNION

SELECT 'SECOND', sum(value)
    FROM  SYSIBM.SYSDUMMY1 GROUP BY 'SECOND' HAVING sum(value) IS NOT NULL

UNION

SELECT 'THIRD', sum(value)
    FROM  SYSIBM.SYSDUMMY1 GROUP BY 'THIRD' HAVING sum(value) IS NOT NULL

如果sum(value)中的任何一个返回NULL，则由于SELECT子句，关联的GROUP BY语句将返回零行，从而导致相应的行省略。 / p>

如果您想省略基于 a NULL 或零和的行，请像这样添加AND连词：< / p>

SELECT 'FIRST', sum(value) FROM SYSIBM.SYSDUMMY1 GROUP BY 'FIRST' HAVING sum(value) IS NOT NULL AND NOT sum(value) = 0
UNION
SELECT 'SECOND', sum(value) FROM  SYSIBM.SYSDUMMY1 GROUP BY 'SECOND' HAVING sum(value) IS NOT NULL AND NOT sum(value) = 0
UNION
SELECT 'THIRD', sum(value) FROM  SYSIBM.SYSDUMMY1 GROUP BY 'THIRD' HAVING sum(value) IS NOT NULL AND NOT sum(value) = 0

此外，您可以通过将UNION更改为UNION ALL来提高查询的效率，因为您的最终结果中不会有重复的行。 UNION执行重复检查，在这种情况下永远不会触发。

SQL DB2 Union跳过NULL / 0行

2 个答案: