㈠ 函数索引的陷阱
使用函数索引一定要注意在函数代码变更后重建函数索引、否则、Oracle将返回错误结果但不给提示
测试如下:
hr@ORCL> drop table t purge;
Table dropped.
hr@ORCL> create table t (x number,y varchar2(30));
Table created.
hr@ORCL> insert into t select rownum,rownum||'a' from dual connect by rownum<1000;
999 rows created.
hr@ORCL> ed
Wrote file afiedt.buf
1 create or replace function f_david(p_value varchar2) return varchar2
2 deterministic is
3 begin
4 return p_value;
5* end;
6
7 /
Function created.
hr@ORCL> create index idx_f_david_t on t (f_david(y));
Index created.
hr@ORCL> exec dbms_stats.gather_table_stats(ownname=>'HR',tabname=>'T',estimate_percent=>100,cascade=>TRUE,no_invalidate=>false);
PL/SQL procedure successfully completed.
hr@ORCL> select * from t where f_david(y)='8a';
X Y
---------- ------------------------------
8 8a
hr@ORCL> ed
Wrote file afiedt.buf
1 create or replace function f_david(p_value varchar2) return varchar2
2 deterministic is
3 begin
4 return p_value||'b';
5* end;
hr@ORCL> /
Function created.
/* 此时的函数 f_david 已经不是我们所认识的那个了、但是查询依然如故!!!*/
hr@ORCL> select * from t where f_david(y)='8a';
X Y
---------- ------------------------------
8 8a
/* 索引重建查询没有记录、这才是我们要的正确结果*/
hr@ORCL> drop index idx_f_david_t;
Index dropped.
hr@ORCL> create index idx_f_david_t on t (f_david(y));
Index created.
hr@ORCL> exec dbms_stats.gather_table_stats(ownname=>'HR',tabname=>'T',estimate_percent=>100,cascade=>TRUE,no_invalidate=>false);
PL/SQL procedure successfully completed.
hr@ORCL> select * from t where f_david(y)='8a';
no rows selected
㈡ 避免索引被污染
这里给出 2 条意见、
① 不要在字段前增加函数
如:
to_char(start_time,'yyyy.mm.dd') between '2013.06.06' and '2013.06.10'
和
start_time between to_date('2013.06.06','yyyy.mm.dd') and to_date('2013.06.10','yyyy.mm.dd')
任何时候都应该是第二种!!!
② 不要把字段嵌入到表达式中
如:
start_time + 7 < sysdate;
和
start_time < sysdate - 7
By David Lin
2013-06-06
Good Luck