外键一定要加索引

Mysql创建外键约束会自动创建一个索引，但是Oracle不会这样做。
所以Oracle在创建外键之后，一定要手动创建索引。
如果没有在外键上创建索引，会经常遇到如下问题。
1.死锁
    

--创建实验表
create table emp as select * from hr.employees;
create table dept as select * from hr.departments;
alter table emp add constraint pk_emp primary key (employee_id);
alter table dept add constraint pk_dept primary key (department_id);
alter table emp add constraint fk_emp_dept foreign key (department_id) references dept(department_id) on delete cascade;

    如果删除任意一条Emp表的记录，都会导致dept表的所有记录不能删除。
    
    SessionA，删除任意一条Emp表记录，不提交

    
    另一个SessionB,他删除任何dept表的记录，都会被阻塞。


    当然，对Dept表的插入和更新不受影响
    (假设不更新主键字段)

2.性能下降
    查询某个部门的员工信息。

select employee_id,dept.department_name,first_name
from dept
inner join emp
on emp.department_id=dept.department_id
where emp.department_id=80;

    如果没有索引,则emp表全表扫描
    
    如果外键建立索引，则优化的效果显而易见
    

下面是Tom给出的脚本，检查外键无索引的表

COLUMN COLUMNS format a30 word_wrapped
COLUMN tablename format a15 word_wrapped
COLUMN constraint_name format a15 word_wrapped
SELECT TABLE_NAME,
       CONSTRAINT_NAME,
       CNAME1 || NVL2(CNAME2, ',' || CNAME2, NULL) ||
       NVL2(CNAME3, ',' || CNAME3, NULL) ||
       NVL2(CNAME4, ',' || CNAME4, NULL) ||
       NVL2(CNAME5, ',' || CNAME5, NULL) ||
       NVL2(CNAME6, ',' || CNAME6, NULL) ||
       NVL2(CNAME7, ',' || CNAME7, NULL) ||
       NVL2(CNAME8, ',' || CNAME8, NULL) COLUMNS
  FROM (SELECT B.TABLE_NAME,
               B.CONSTRAINT_NAME,
               MAX(DECODE(POSITION, 1, COLUMN_NAME, NULL)) CNAME1,
               MAX(DECODE(POSITION, 2, COLUMN_NAME, NULL)) CNAME2,
               MAX(DECODE(POSITION, 3, COLUMN_NAME, NULL)) CNAME3,
               MAX(DECODE(POSITION, 4, COLUMN_NAME, NULL)) CNAME4,
               MAX(DECODE(POSITION, 5, COLUMN_NAME, NULL)) CNAME5,
               MAX(DECODE(POSITION, 6, COLUMN_NAME, NULL)) CNAME6,
               MAX(DECODE(POSITION, 7, COLUMN_NAME, NULL)) CNAME7,
               MAX(DECODE(POSITION, 8, COLUMN_NAME, NULL)) CNAME8,
               COUNT(*) COL_CNT
          FROM (SELECT SUBSTR(TABLE_NAME, 1, 30) TABLE_NAME,
                       SUBSTR(CONSTRAINT_NAME, 1, 30) CONSTRAINT_NAME,
                       SUBSTR(COLUMN_NAME, 1, 30) COLUMN_NAME,
                       POSITION
                  FROM USER_CONS_COLUMNS) A,
               USER_CONSTRAINTS B
         WHERE A.CONSTRAINT_NAME = B.CONSTRAINT_NAME
           AND B.CONSTRAINT_TYPE = 'R'
         GROUP BY B.TABLE_NAME, B.CONSTRAINT_NAME) CONS
 WHERE COL_CNT > ALL
 (SELECT COUNT(*)
          FROM USER_IND_COLUMNS I
         WHERE I.TABLE_NAME = CONS.TABLE_NAME
           AND I.COLUMN_NAME IN (CNAME1, CNAME2, CNAME3, CNAME4, CNAME5,
                CNAME6, CNAME7, CNAME8)
           AND I.COLUMN_POSITION <= CONS.COL_CNT
         GROUP BY I.INDEX_NAME)
/

参考：
http://blog.csdn.net/dba_waterbin/article/details/9045483

来自 “ ITPUB博客 ” ，链接：http://blog.itpub.net/29254281/viewspace-1065217/，如需转载，请注明出处，否则将追究法律责任。

转载于:http://blog.itpub.net/29254281/viewspace-1065217/