HINT如何跨越視圖生效 .txt
前兩天同事問我這個問題，估計不少人都不清楚，順便寫出來。
Oracle的HINT能夠強制指定SQL的執行計劃，好比選擇索引、表的鏈接順序以及表的鏈接方式等等。
不過若是查詢的是視圖，使用HINT就變得麻煩一些。
看一個具體的例子吧：
SQL> CREATE TABLE T1 (ID NUMBER, NAME VARCHAR2(30), TYPE VARCHAR2(30));
表已建立。
SQL> CREATE TABLE T2 (ID NUMBER, NAME VARCHAR2(30), TYPE VARCHAR2(30));
表已建立。
SQL> CREATE INDEX IND_T1_NAME ON T1 (NAME);
索引已建立。
SQL> CREATE INDEX IND_T2_NAME ON T2 (NAME);
索引已建立。
SQL> INSERT INTO T1 SELECT ROWNUM, OBJECT_NAME, OBJECT_TYPE
FROM DBA_OBJECTS WHERE OBJECT_TYPE = 'TABLE';
已建立2058行。
SQL> INSERT INTO T2 SELECT ROWNUM, OBJECT_NAME, OBJECT_TYPE
FROM DBA_OBJECTS WHERE OBJECT_TYPE = 'TABLE';
已建立2058行。
SQL> COMMIT;
提交完成。
SQL> CREATE VIEW V_T AS
SELECT * FROM T1 UNION ALL SELECT * FROM T2;
視圖已建立。
建立了一個包含UNION ALL查詢的視圖，下面執行對這個視圖的查詢：
SQL> SET AUTOT ON EXP
SQL> SELECT *
FROM V_T WHERE NAME LIKE 'B%';
ID NAME TYPE
---------- ------------------------------ ------------------------------
13 BOOTSTRAP$ TABLE
1774 BONUS TABLE
1871 BS TABLE
1872 BP TABLE
1873 BCF TABLE
1876 BSF TABLE
1877 BDF TABLE
1880 BRL TABLE
1881 BCB TABLE
13 BOOTSTRAP$ TABLE
1774 BONUS TABLE
1871 BS TABLE
1872 BP TABLE
1873 BCF TABLE
1876 BSF TABLE
1877 BDF TABLE
1880 BRL TABLE
1881 BCB TABLE
已選擇18行。
執行計劃
----------------------------------------------------------
Plan hash value: 2128801320
-----------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
-----------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 55 | 2585 | 2 (0)| 00:00:01 |
| 1 | VIEW | V_T | 55 | 2585 | 2 (0)| 00:00:01 |
| 2 | UNION-ALL PARTITION| | | | | |
|* 3 | TABLE ACCESS FULL | T1 | 9 | 423 | 5 (0)| 00:00:01 |
|* 4 | TABLE ACCESS FULL | T2 | 9 | 423 | 5 (0)| 00:00:01 |
-----------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
3 - filter("NAME" LIKE 'B%')
4 - filter("NAME" LIKE 'B%')
Note
-----
- dynamic sampling used for this statement
能夠看到，Oracle沒有使用索引，對於這個SQL來講，若是T1和T2都使用索引掃描的話，效率會更高，
可是，對於包含視圖的查詢，普通的HINT指定方法是無效的：
SQL> SELECT /*+ INDEX(T1 IND_T1_NAME) */ *
FROM V_T WHERE NAME LIKE 'B%';
ID NAME TYPE
---------- ------------------------------ ------------------------------
13 BOOTSTRAP$ TABLE
1774 BONUS TABLE
1871 BS TABLE
1872 BP TABLE
1873 BCF TABLE
1876 BSF TABLE
1877 BDF TABLE
1880 BRL TABLE
1881 BCB TABLE
13 BOOTSTRAP$ TABLE
1774 BONUS TABLE
1871 BS TABLE
1872 BP TABLE
1873 BCF TABLE
1876 BSF TABLE
1877 BDF TABLE
1880 BRL TABLE
1881 BCB TABLE
已選擇18行。
執行計劃
----------------------------------------------------------
Plan hash value: 2128801320
-----------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
-----------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 55 | 2585 | 2 (0)| 00:00:01 |
| 1 | VIEW | V_T | 55 | 2585 | 2 (0)| 00:00:01 |
| 2 | UNION-ALL PARTITION| | | | | |
|* 3 | TABLE ACCESS FULL | T1 | 9 | 423 | 5 (0)| 00:00:01 |
|* 4 | TABLE ACCESS FULL | T2 | 9 | 423 | 5 (0)| 00:00:01 |
-----------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
3 - filter("NAME" LIKE 'B%')
4 - filter("NAME" LIKE 'B%')
Note
-----
- dynamic sampling used for this statement
其實指定HINT的方法並不複雜，只須要指定VIEW的名稱做爲表名的前綴就能夠了：
SQL> SELECT /*+ INDEX(V_T.T1 IND_T1_NAME) */ *
FROM V_T WHERE NAME LIKE 'B%';
ID NAME TYPE
---------- ------------------------------ ------------------------------
1881 BCB TABLE
1873 BCF TABLE
1877 BDF TABLE
1774 BONUS TABLE
13 BOOTSTRAP$ TABLE
1872 BP TABLE
1880 BRL TABLE
1871 BS TABLE
1876 BSF TABLE
13 BOOTSTRAP$ TABLE
1774 BONUS TABLE
1871 BS TABLE
1872 BP TABLE
1873 BCF TABLE
1876 BSF TABLE
1877 BDF TABLE
1880 BRL TABLE
1881 BCB TABLE
已選擇18行。
執行計劃
----------------------------------------------------------
Plan hash value: 1398832356
--------------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
--------------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 55 | 2585 | 2 (0)| 00:00:01 |
| 1 | VIEW | V_T | 55 | 2585 | 2 (0)| 00:00:01 |
| 2 | UNION-ALL PARTITION | | | | | |
| 3 | TABLE ACCESS BY INDEX ROWID| T1 | 9 | 423 | 6 (0)| 00:00:01 |
|* 4 | INDEX RANGE SCAN | IND_T1_NAME | 9 | | 2 (0)| 00:00:01 |
|* 5 | TABLE ACCESS FULL | T2 | 9 | 423 | 5 (0)| 00:00:01 |
--------------------------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
4 - access("NAME" LIKE 'B%')
filter("NAME" LIKE 'B%')
5 - filter("NAME" LIKE 'B%')
Note
-----
- dynamic sampling used for this statement
一樣對於T2表也能夠採用一樣的方法：
SQL> SELECT /*+ INDEX(V_T.T1 IND_T1_NAME) INDEX(V_T.T2 IND_T2_NAME) */ *
FROM V_T WHERE NAME LIKE 'B%';
ID NAME TYPE
---------- ------------------------------ ------------------------------
1881 BCB TABLE
1873 BCF TABLE
1877 BDF TABLE
1774 BONUS TABLE
13 BOOTSTRAP$ TABLE
1872 BP TABLE
1880 BRL TABLE
1871 BS TABLE
1876 BSF TABLE
1881 BCB TABLE
1873 BCF TABLE
1877 BDF TABLE
1774 BONUS TABLE
13 BOOTSTRAP$ TABLE
1872 BP TABLE
1880 BRL TABLE
1871 BS TABLE
1876 BSF TABLE
已選擇18行。
執行計劃
----------------------------------------------------------
Plan hash value: 4053622206
--------------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
--------------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 55 | 2585 | 2 (0)| 00:00:01 |
| 1 | VIEW | V_T | 55 | 2585 | 2 (0)| 00:00:01 |
| 2 | UNION-ALL PARTITION | | | | | |
| 3 | TABLE ACCESS BY INDEX ROWID| T1 | 9 | 423 | 6 (0)| 00:00:01 |
|* 4 | INDEX RANGE SCAN | IND_T1_NAME | 9 | | 2 (0)| 00:00:01 |
| 5 | TABLE ACCESS BY INDEX ROWID| T2 | 9 | 423 | 6 (0)| 00:00:01 |
|* 6 | INDEX RANGE SCAN | IND_T2_NAME | 9 | | 2 (0)| 00:00:01 |
--------------------------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
4 - access("NAME" LIKE 'B%')
filter("NAME" LIKE 'B%')
6 - access("NAME" LIKE 'B%')
filter("NAME" LIKE 'B%')
Note
-----
- dynamic sampling used for this statement
若是視圖中使用了別名，那麼HINT中指定的時候也要使用別名：
SQL> CREATE OR REPLACE VIEW V_T AS
SELECT * FROM T1 A UNION ALL SELECT * FROM T2 B;
視圖已建立。
SQL> SELECT /*+ INDEX(V_T.T1 IND_T1_NAME) INDEX(V_T.T2 IND_T2_NAME) */ *
FROM V_T WHERE NAME LIKE 'B%';
ID NAME TYPE
---------- ------------------------------ ------------------------------
13 BOOTSTRAP$ TABLE
1774 BONUS TABLE
1871 BS TABLE
1872 BP TABLE
1873 BCF TABLE
1876 BSF TABLE
1877 BDF TABLE
1880 BRL TABLE
1881 BCB TABLE
13 BOOTSTRAP$ TABLE
1774 BONUS TABLE
1871 BS TABLE
1872 BP TABLE
1873 BCF TABLE
1876 BSF TABLE
1877 BDF TABLE
1880 BRL TABLE
1881 BCB TABLE
已選擇18行。
執行計劃
----------------------------------------------------------
Plan hash value: 2128801320
-----------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
-----------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 55 | 2585 | 2 (0)| 00:00:01 |
| 1 | VIEW | V_T | 55 | 2585 | 2 (0)| 00:00:01 |
| 2 | UNION-ALL PARTITION| | | | | |
|* 3 | TABLE ACCESS FULL | T1 | 9 | 423 | 5 (0)| 00:00:01 |
|* 4 | TABLE ACCESS FULL | T2 | 9 | 423 | 5 (0)| 00:00:01 |
-----------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
3 - filter("NAME" LIKE 'B%')
4 - filter("NAME" LIKE 'B%')
Note
-----
- dynamic sampling used for this statement
SQL> SELECT /*+ INDEX(V_T.A IND_T1_NAME) INDEX(V_T.B IND_T2_NAME) */ *
FROM V_T WHERE NAME LIKE 'B%';
ID NAME TYPE
---------- ------------------------------ ------------------------------
1881 BCB TABLE
1873 BCF TABLE
1877 BDF TABLE
1774 BONUS TABLE
13 BOOTSTRAP$ TABLE
1872 BP TABLE
1880 BRL TABLE
1871 BS TABLE
1876 BSF TABLE
1881 BCB TABLE
1873 BCF TABLE
1877 BDF TABLE
1774 BONUS TABLE
13 BOOTSTRAP$ TABLE
1872 BP TABLE
1880 BRL TABLE
1871 BS TABLE
1876 BSF TABLE
已選擇18行。
執行計劃
----------------------------------------------------------
Plan hash value: 4053622206
--------------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
--------------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 55 | 2585 | 2 (0)| 00:00:01 |
| 1 | VIEW | V_T | 55 | 2585 | 2 (0)| 00:00:01 |
| 2 | UNION-ALL PARTITION | | | | | |
| 3 | TABLE ACCESS BY INDEX ROWID| T1 | 9 | 423 | 6 (0)| 00:00:01 |
|* 4 | INDEX RANGE SCAN | IND_T1_NAME | 9 | | 2 (0)| 00:00:01 |
| 5 | TABLE ACCESS BY INDEX ROWID| T2 | 9 | 423 | 6 (0)| 00:00:01 |
|* 6 | INDEX RANGE SCAN | IND_T2_NAME | 9 | | 2 (0)| 00:00:01 |
--------------------------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
4 - access("NAME" LIKE 'B%')
filter("NAME" LIKE 'B%')
6 - access("NAME" LIKE 'B%')
filter("NAME" LIKE 'B%')
Note
-----
- dynamic sampling used for this statement
相關文章
- 1. DB2如何創建跨庫視圖
- 2. oss如何解決跨越問題
- 3. 從切圖仔到前端 Leader，如何跨越式成長
- 4. 如何在UI5視圖裏包含原生的HTML視圖
- 5. 原生js，實現跨越的jsonp
- 6. SOLIDWORKS如何簡單高效地自動生成模型視圖和標註
- 7. 跨項目視圖
- 8. oracle如何建立視圖
- 9. ABAP中如何建視圖
- 10. 【跨團隊協做】如何有效的跨團隊協做
- 更多相關文章...
- • XSD 如何使用? - XML Schema 教程
- • 如何僞造ARP響應？ - TCP/IP教程
- • PHP Ajax 跨域問題最佳解決方案
- • Git可視化極簡易教程 — Git GUI使用方法
相關標籤/搜索
每日一句
-
每一个你不满意的现在，都有一个你没有努力的曾经。
歡迎關注本站公眾號,獲取更多信息
相關文章
- 1. DB2如何創建跨庫視圖
- 2. oss如何解決跨越問題
- 3. 從切圖仔到前端 Leader，如何跨越式成長
- 4. 如何在UI5視圖裏包含原生的HTML視圖
- 5. 原生js，實現跨越的jsonp
- 6. SOLIDWORKS如何簡單高效地自動生成模型視圖和標註
- 7. 跨項目視圖
- 8. oracle如何建立視圖
- 9. ABAP中如何建視圖
- 10. 【跨團隊協做】如何有效的跨團隊協做