Question

我有这张桌子：

CREATE TABLE "TEST"."TEST_REQUESTS"
  (
    "REQUEST_ID" NUMBER(*,0) NOT NULL ENABLE,
    "RECEPTION_TIME" TIMESTAMP (6) NOT NULL ENABLE,
    "OPERATION" VARCHAR2(25 BYTE) NOT NULL ENABLE,
    "XML_HEADER_IN" CLOB,
    "XML_BODY_IN" CLOB NOT NULL ENABLE,
    "STATUS_ID" NUMBER(*,0) NOT NULL ENABLE,
    "CUSTOMER"  VARCHAR2(25 BYTE)
  )

它有2个索引：

CREATE INDEX "TEST"."TEST_REQUESTS_STATUS_IDX" ON "TEST"."TEST_REQUESTS"
    (
      "STATUS_ID" DESC
    )
CREATE INDEX "TEST"."TEST_REQUESTS_TIME_IDX" ON "TEST"."TEST_REQUESTS"
    (
      "RECEPTION_TIME"
    )

我也有一个或多或少的程序：

PROCEDURE TEST_LoadHistoryRequestRange (diasMantenidosIN IN number, [...] numRequestOut OUT number [...]) IS

tsStartLimit TIMESTAMP;

BEGIN

tsStartLimit := sysdate - diasMantenidosIN;

SELECT count(REQUEST_ID) into numRequestOut FROM TEST_REQUESTS WHERE STATUS_ID=0 and RECEPTION_TIME<tsStartLimit;

[...]

END;

现在这个表已经增长到接近1500万个寄存器......但是，花费近10分钟来制作那么简单COUNT是合乎逻辑的吗？这可能有什么问题？

非常感谢!!

编辑：这是选择计数的解释计划：

Plan hash value: 378939817

----------------------------------------------------------------------------------------------------------
| Id  | Operation                    | Name                      | Rows  | Bytes | Cost (%CPU)| Time     |
----------------------------------------------------------------------------------------------------------
|   0 | SELECT STATEMENT             |                           |     1 |    14 |     4   (0)| 00:00:01 |
|   1 |  SORT AGGREGATE              |                           |     1 |    14 |            |          |
|*  2 |   TABLE ACCESS BY INDEX ROWID| TEST_REQUESTS             |  1125 | 15750 |     4   (0)| 00:00:01 |
|*  3 |    INDEX RANGE SCAN          | TEST_REQUESTS_STATUS_IDX  |     1 |       |     3   (0)| 00:00:01 |
----------------------------------------------------------------------------------------------------------

Answer 1

可能需要很长时间。查询的最佳索引是复合索引：

CREATE INDEX "TEST"."TEST_STATUS_REQUESTS_TIME_IDX" ON "TEST"."TEST_REQUESTS"
    (
      STATUS,
      "RECEPTION_TIME",
      REQUEST_ID
    );

请注意，索引中列的顺序会有所不同。

Oracle Timestamp索引非常慢

1 个答案: