在使用implict update table 时,发现其中一个表一定要有唯一约束,否则会报错!但是oracle可以使用hints:/*+ BYPASS_UJVC*/ 屏蔽掉队唯一性的检查。具体测试过程如下:
SQL> CREATE TABLE test_a(
2 id NUMBER ,
3 score NUMBER );
Table created.
Elapsed: 00:00:00.10
SQL>
CREATE TABLE test_b(
2 id NUMBER);
Table created.
Elapsed: 00:00:00.01
SQL>
SQL> INSERT INTO test_a VALUES(1,100);
INSERT INTO test_a VALUES(2,200);
1 row created.
Elapsed: 00:00:00.01
SQL>
INSERT INTO test_a VALUES(3,300);
1 row created.
Elapsed: 00:00:00.00
SQL>
1 row created.
Elapsed: 00:00:00.00
SQL> INSERT INTO test_b VALUES(1);
1 row created.
Elapsed: 00:00:00.01
SQL> INSERT INTO test_b VALUES(2);
1 row created.
Elapsed: 00:00:00.00
SQL> INSERT INTO test_b VALUES(3);
1 row created.
Elapsed: 00:00:00.01
SQL> INSERT INTO test_b VALUES(4);
1 row created.
Elapsed: 00:00:00.01
SQL>
SQL> SELECT * FROM test_a ORDER BY 1;
SELECT * FROM test_b ORDER BY 1;
ID SCORE
---------- ----------
1 100
2 200
3 300
3 rows selected.
Elapsed: 00:00:00.01
SQL>
ID
----------
1
2
3
4
4 rows selected.
Elapsed: 00:00:00.01
现在更新a表的字段:(先不考虑SQL的功能,只是测试这种方法)
SQL> UPDATE (
2 SELECT a.id,a.score,b.id AS b_id
3 FROM test_a a,
4 test_b b
5 WHERE a.id=b.id
6 )
7 SET id=b_id;
SET id=b_id
*
ERROR at line 7:
ORA-01779: cannot modify a column which maps to a non key-preserved table
报错了,需要对b表加唯一索引。加入hints执行
SQL> UPDATE (
2 SELECT/*+ BYPASS_UJVC*/ a.id,a.score,b.id AS b_id
3 FROM test_a a,
4 test_b b
5 WHERE a.id=b.id
6 )
7 SET id=b_id;
3 rows updated.
Elapsed: 00:00:00.01
说明:oracle可以跳过检查唯一约束。
继续往b表添加一条记录,使b表的记录不唯一。
SQL> INSERT INTO test_b VALUES(3);
1 row created.
Elapsed: 00:00:00.00
SQL> COMMIT;
Commit complete.
Elapsed: 00:00:00.00
SQL> SELECT * FROM TEST_B ORDER BY 1;
ID
----------
1
2
3
3
4
5 rows selected.
Elapsed: 00:00:00.00
SQL> UPDATE (
2 SELECT/*+ BYPASS_UJVC*/ a.id,a.score,b.id AS b_id
3 FROM test_a a,
4 test_b b
5 WHERE a.id=b.id
6 )
7 SET id=b_id;
4 rows updated.
Elapsed: 00:00:00.01
SQL> SELECT * FROM TEST_A;
ID SCORE
---------- ----------
1 100
2 200
3 300
3 rows selected.
Elapsed: 00:00:00.00
总结:使用了该hints,oracle就完全放弃了检查唯一性。因此SQL可以执行通过。
但要注意,是否能达到我们希望的目的,例如:下面进行一个有意义的更新,将a表中只要id在b表中出现就更新score,将score加1.
执行SQL如下:
SQL> UPDATE (
2 SELECT/*+ BYPASS_UJVC*/ a.id,a.score,b.id AS b_id
3 FROM test_a a,
4 test_b b
5 WHERE a.id=b.id
6 )
7 SET score=score+1;
4 rows updated.
Elapsed: 00:00:00.00
SQL> SELECT * FROM TEST_A;
ID SCORE
---------- ----------
1 101
2 201
3 302
3 rows selected.
Elapsed: 00:00:00.01
发现对于id=3的记录多加了个1。因此并没有得到我们想要的结果。这种情况应该先对b表的id进行去重。然后再更新。(其实完全也可以换种写法,呵呵) ,在此就不写了
总结:虽然对于此种应用,oracle跳过了检查唯一性,但是我们要注意在具体使用时,是否真的就达到了我们的效果!
要巧用并且活用现在已经有的功能。