我正在使用以下查询在 Oracle 11g(11.2.0.3.0)中创建视图。
CREATE OR REPLACE FORCE VIEW V_DOCUMENTS_LIST
(
ID_DOC,
ATTACHMENTS_COUNT,
TOTAL_DIMENSION,
INSERT_DATE,
ID_STATE,
STATE,
ID_INSTITUTE,
INSTITUTE,
HASJOB
)
AS
SELECT D.ID_DOC,
COUNT (F.ID_FILE) AS ATTACHMENTS_COUNT,
CASE
WHEN SUM (F.DIMENSION) IS NULL THEN 0
ELSE SUM (F.DIMENSION)
END
AS TOTAL_DIMENSION,
D.INSERT_DATE,
D.ID_STATE,
S.STATE_DESC AS STATE,
D.ID_INSTITUTE,
E.NAME AS INSTITUTE,
CASE
WHEN EXISTS (SELECT D.ID_DOC FROM JOB) THEN 'true'
ELSE 'false'
END
AS HASJOB
FROM DOCUMENTS D
LEFT JOIN FILES F ON D.ID_DOC = F.ID_DOC
JOIN STATES S ON D.ID_STATE = S.ID_STATE
JOIN INSTITUTES E ON D.ID_INSTITUTE = E.ID_INSTITUTE
GROUP BY D.ID_DOC,
D.INSERT_DATE,
D.ID_STATE,
S.STATE_DESC,
D.ID_INSTITUTE,
E.NAME;
然后我查询该视图以获取DataGridView
页面中ASPX
的值。
SELECT *
FROM V_DOCUMENTS_LIST
ORDER BY ID_STATE DESC, INSTITUTE, INSERT_DATE DESC;
相关表和关系
文件;文件;的工作;
文件(1-1)< ----> (0-N)文件
JOBS(0-1)< ----> (0-N)文件
查询视图我得到完整的文档列表及其所有相关信息(ID,描述,日期,状态等)以及每个文档:
JOB
与 DOCUMENT
相关联> 。一切都工作得很好,包含几千条记录的视图。现在记录数量正在增加,视图上的SELECT * FROM
大约需要2:30分钟,记录为15.000-20.000。
我知道我的视图中非常耗时的部分是嵌套的SELECT
:
CASE
WHEN EXISTS (SELECT D.ID_DOC FROM JOB) THEN 'true'
ELSE 'false'
END
AS HASJOB
如何优化观点?
答案 0 :(得分:4)
要解决不存在的问题,您可以添加联接:
LEFT JOIN (select distinct id_doc from JOB) J
ON d.id_doc = J.id_doc
Has_job列将是:
CASE
WHEN j.id_doc is not null THEN 'true'
ELSE 'false'
END AS HASJOB
PS :您当前的实现存在问题,因为如果作业表有行,SELECT D.ID_DOC FROM JOB
将始终包含行。它等同于select * from job
,因为只存在行的测试存在。逻辑上正确的实现方式是:SELECT 1 FROM JOB j where j.id_doc = D.ID_DOC
。
答案 1 :(得分:0)
您将在表JOB
上使用完整索引,在查询中添加WHERE
子句:
SELECT D.ID_DOC FROM JOB