从 SQL 查询 (Oracle) 的字段中的最小值获取所有行数据

Question

我无法通过指定字段的最小值从 SQL 查询中获取所有信息的行。下面是我正在使用的数据以及我想要获取的数据的基本示例：

SELECT 1 AS NUM_, 'ABC' AS LET_ FROM DUAL
UNION
SELECT 2 AS NUM_, 'DEF' AS LET_ FROM DUAL
UNION
SELECT 3 AS NUM_, 'GHI' AS LET_ FROM DUAL;

上面的查询将产生以下结果：

NUM_ | LET_
-----------
  1  | ABC
  2  | DEF
  3  | GHI

我只想要包含数据1 和ABC 的行。这是我尝试过的，如下：

SELECT MIN(LN.NUM_) AS MIN_NUM, 
       LN.LET_ 
FROM   (SELECT 1 AS NUM_, 'ABC' AS LET_ FROM DUAL
        UNION
        SELECT 2 AS NUM_, 'DEF' AS LET_ FROM DUAL
        UNION
        SELECT 3 AS NUM_, 'GHI' AS LET_ FROM DUAL) LN
GROUP BY LET_;

但上面仍然给了我所有的行。如何仅将NUM_ 列中值最低的行归零，给我1 和ABC？

Answer 1

您的查询会找到每个 LET_ 的最小值 NUM_。要在所有记录中获得最小的NUM_，请使用ORDER BY & ROWNUM

试试这个

SELECT LN.NUM_ AS MIN_NUM, 
       LN.LET_ 
FROM   (SELECT 1 AS NUM_, 'ABC' AS LET_ FROM DUAL
        UNION
        SELECT 2 AS NUM_, 'DEF' AS LET_ FROM DUAL
        UNION
        SELECT 3 AS NUM_, 'GHI' AS LET_ FROM DUAL
        ORDER BY LN.NUM_) LN
Where ROWNUM = 1

Answer 2

在 Oracle 12C 中，您可以使用 ANSI 标准 fetch first row only 语法：

SELECT 1 AS NUM_, 'ABC' AS LET_ FROM DUAL
UNION ALL
SELECT 2 AS NUM_, 'DEF' AS LET_ FROM DUAL
UNION ALL
SELECT 3 AS NUM_, 'GHI' AS LET_ FROM DUAL
ORDER BY 1
FETCH FIRST 1 ROW ONLY;

Answer 3

正如已经指出的，有不同的方法可以做到这一点：

一些方法：

• FETCH FIRST 1 ROW ONLY
• ORDER BY 在视图中然后 ROWNUM = 1
• （我的补充）MAX KEEP
• （我的补充）不相关子查询

FETCH FIRST 1 ROW ONLY 是最简单的语法，并且性能足够好。但如果你有一张大桌子，其他方法也值得考虑。

您的数据集太小，无法展示不同的性能结果，所以为了说明它们，让我们用DBA_OBJECTS 创建一个表格。

create table matt_test as SELECT * FROM dba_objects;
create unique index matt_test_u1 on matt_test (object_id);
exec dbms_stats.gather_table_stats(USER, 'MATT_TEST');

然后，尝试一些不同的方法并使用 DBMS_XPLAN 进行测量：

最大保持

SELECT MAX (object_id) KEEP (DENSE_RANK FIRST ORDER BY object_id) object_id,
       MAX (object_name) KEEP (DENSE_RANK FIRST ORDER BY object_id) object_name
FROM   matt_test o;

------------------------------------------------------------------------------------------
| Id  | Operation          | Name      | Starts | E-Rows | A-Rows |   A-Time   | Buffers |
------------------------------------------------------------------------------------------
|   0 | SELECT STATEMENT   |           |      1 |        |      1 |00:00:00.29 |    8522 |
|   1 |  SORT AGGREGATE    |           |      1 |      1 |      1 |00:00:00.29 |    8522 |
|   2 |   TABLE ACCESS FULL| MATT_TEST |      1 |    555K|    555K|00:00:00.12 |    8522 |
------------------------------------------------------------------------------------------

FETCH FIRST（注意内存使用）

SELECT object_id, object_name
FROM   matt_test o
order by object_id
fetch first 1 rows only;

---------------------------------------------------------------------------------------------------------------------------
| Id  | Operation                | Name      | Starts | E-Rows | A-Rows |   A-Time   | Buffers |  OMem |  1Mem | Used-Mem |
---------------------------------------------------------------------------------------------------------------------------
|   0 | SELECT STATEMENT         |           |      1 |        |      1 |00:00:00.33 |    8522 |       |       |          |
|*  1 |  VIEW                    |           |      1 |      1 |      1 |00:00:00.33 |    8522 |       |       |          |
|*  2 |   WINDOW SORT PUSHED RANK|           |      1 |    555K|      1 |00:00:00.33 |    8522 |  2048 |  2048 | 2048  (0)|
|   3 |    TABLE ACCESS FULL     | MATT_TEST |      1 |    555K|    555K|00:00:00.12 |    8522 |       |       |          |
---------------------------------------------------------------------------------------------------------------------------

不相关的子查询

select object_id, object_name
from matt_test
where object_id = ( SELECT min(object_id) FROM matt_test );

-------------------------------------------------------------------------------------------------------
| Id  | Operation                    | Name         | Starts | E-Rows | A-Rows |   A-Time   | Buffers |
-------------------------------------------------------------------------------------------------------
|   0 | SELECT STATEMENT             |              |      1 |        |      1 |00:00:00.01 |       7 |
|   1 |  TABLE ACCESS BY INDEX ROWID | MATT_TEST    |      1 |      1 |      1 |00:00:00.01 |       7 |
|*  2 |   INDEX UNIQUE SCAN          | MATT_TEST_U1 |      1 |      1 |      1 |00:00:00.01 |       6 |
|   3 |    SORT AGGREGATE            |              |      1 |      1 |      1 |00:00:00.01 |       3 |
|   4 |     INDEX FULL SCAN (MIN/MAX)| MATT_TEST_U1 |      1 |      1 |      1 |00:00:00.01 |       3 |
-------------------------------------------------------------------------------------------------------

Predicate Information (identified by operation id):
---------------------------------------------------

   2 - access("OBJECT_ID"=)

因此，如果您可以利用索引，您可以看到不相关的子查询可以快得多。 MAX KEEP 方法的性能比FETCH FIRST ROWS 方法好一点，因为它使用的内存更少。

没有一种最好的方法：每一种都有它的位置。

如果我在不考虑性能的情况下编写此代码（例如，小型数据集），则默认方法是 FETCH FIRST ROWS。