Not used for subsequent inserts
(先建索引,再插入数据(不受pctfree的影响)。与先插入数据,再建索引(受pctfree的影响)。这两种情 况对空间使用完全不同)
例如:先插入数据,再建索引
CREATE TABLE t1 (c01 NUMBER);
Insert 100000 rows
CREATE INDEX i1 ON t1 (c01)
PCTFREE 50
先建索引,再插入数据
CREATE TABLE t1 (c01 NUMBER);
CREATE INDEX i1 ON t1 (c01)
PCTFREE 50;
Insert 100000 rows
试验:先插入数据,再建索引
SQL> create table t6(c1 number);
Table created.
1 declare
2 c number;
3 begin
4 for i in 1..100000 loop
5 c:=i;
6 insert into t6 values(c);
7 end loop;
8 commit;
9* end;
SQL> /
PL/SQL procedure successfully completed.
SQL> create index t6_ind on t6(c1)
2 pctfree 50;
Index created.
SQL> analyze table t6 compute statistics for all indexes;
Table analyzed.
SQL> select * from dba_indexes
2 where table_name='T6';
BLEVEL LEAF_BLOCKS DISTINCT_KEYS
------ ----------- -------------
1 408 100000
先建索引,再插入数据
SQL> create table t7(c1 number);
SQL> create index t7_ind on t6(c1)
2 pctfree 50;
1 declare
2 c number;
3 begin
4 for i in 1..100000 loop
5 c:=i;
6 insert into t6 values(c);
7 end loop;
8 commit;
9* end;
SQL> /
analyze table t7 compute statistics for all indexes;
SQL> select * from dba_indexes
2 where table_name='T7';
BLEVEL LEAF_BLOCKS DISTINCT_KEYS
------- ----------- -------------
1 199 100000
可以看到第二种情况,LEAF_BLOCKS 减少了一半还要多。
来自 “ ITPUB博客 ” ,链接:http://blog.itpub.net/9599/viewspace-472970/,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/9599/viewspace-472970/