顾名思义,基于函数的索引就把要做索引的字段的值先用function计算好,然后对计算好后的值作为索引的值再创建索引,广泛的用在很多场合,比如某个字段的值查询时大小写不敏感等可用此类型索引!当我们创建基于函数的索引时,oracle究竟做了哪些处理我们值得研究一下:
1.首先我们创建测试表:
create table test_fi as select * from all_objects;
select obj# from obj$ where name='TEST_FI';
sys@ORCL> select name, col#, type#,intcol# from col$ where obj#=123368;
NAME COL# TYPE# INTCOL#
------------------------------ ---------- ---------- ----------
GENERATED 12 1 12
SECONDARY 13 1 13
OWNER 1 1 1
OBJECT_NAME 2 1 2
SUBOBJECT_NAME 3 1 3
OBJECT_ID 4 2 4
DATA_OBJECT_ID 5 2 5
OBJECT_TYPE 6 1 6
CREATED 7 12 7
LAST_DDL_TIME 8 12 8
TIMESTAMP 9 1 9
STATUS 10 1 10
TEMPORARY 11 1 11
已选择13行。
sys@ORCL> SELECT cols, intcols, kernelcols FROM tab$
2 WHERE obj# =123368;
COLS INTCOLS KERNELCOLS
---------- ---------- ----------
13 13 13
2:我们创建基于函数的索引,看基表发生哪些变化:
sys@ORCL> create index test_fi_idx on test_fi(upper(owner));
索引已创建。
sys@ORCL> select name, col#, type#,intcol# from col$ where obj#=123368;
NAME COL# TYPE# INTCOL#
------------------------------ ---------- ---------- ----------
GENERATED 12 1 12
SECONDARY 13 1 13
SYS_NC00014$ 0 1 14
OWNER 1 1 1
OBJECT_NAME 2 1 2
SUBOBJECT_NAME 3 1 3
OBJECT_ID 4 2 4
DATA_OBJECT_ID 5 2 5
OBJECT_TYPE 6 1 6
CREATED 7 12 7
LAST_DDL_TIME 8 12 8
TIMESTAMP 9 1 9
STATUS 10 1 10
TEMPORARY 11 1 11
已选择14行
sys@ORCL> SELECT cols, intcols, kernelcols FROM tab$
2 WHERE obj# =123368;
COLS INTCOLS KERNELCOLS
---------- ---------- ----------
13 14 13
这个时候我们发现oracle在col$基表中增加了一列,自动命名,对用户来讲伪列,不可见的,实际上对oracle来讲增加此列是为了统一管理的需要。
sys@ORCL> select index_name, column_name, descend
2 from user_ind_columns where table_name='TEST_FI';
INDEX_NAME COLUMN_NAME DESC
------------------------------ ------------------------------ ----
TEST_FI_IDX SYS_NC00014$ ASC