row-level dependency tracking (10g new feature) SCN_TO_TIMESTAMP ora_rowscn
row-level dependency tracking (10g new feature)SCN_TO_TIMESTAMP ora_rowscn SELECT ORA_ROWSCN,rowid from test;select SCN_TO_TIMESTAMP(205810284) from dual;refs:http://www.dbafan.com/blog/?p=194[@more@]
分区--复合分区
--测试总结:复合分区,有个很奇妙的地方,就是当数据插入到符合条件的第一个分区中。实现算法推测:对于按C1,C2两列进行复合分区的情况,插入时,首先判断是否小于C1,如果有匹配分区,则插入到符合条件的第一个分区中,否则如果插入的值=C1,则判断C2,如果有匹配分区,则插入到符合条件的第一个分区中。第二种理解:复合分区其实等同于单列分区,只是将多个列按二进制存放格式拼组在一起,组成一个单列字段。插入数据时,将分区字段值也按二进制存放格式拼组在一起进行单值判断。drop table pt1 ;create table pt1 (id1 number ,id2 n
cluster table
cluster table :1, if one boundle can acrosss block ? no.2, if each member table exists in user_segments view ? Aso whether all member table take as an independent table ? answer: not exists.3, direct load performance poor ? yes, cluster table Not support direct load .4,[@more@]
truncate操作导致DATA_OBJECT_ID改变
refs:http://www.linuxidc.com/Linux/2011-04/34164.htm 经实验证实,确实如此,比较吃惊。test platform: windows 7 64bit with oracle 11.2.0.l 64bit.[@more@]
constraint deferrable immediate check
-- define a deferrable constraint allow check constraint when a transaction is commited, by default with immediate constraint allow check constraint after DML operation .-- allow primary key check deferrable.alter table t1 add constraint pk_t1 primary key (id) initially deferred ;insert into t1 (id) values (1) ;insert into t1 (id) values (1) ;commit ; -- report error here.-- allow foreign key check deferrable.alter table t2 add constraint fk_t2_t1 foreign key (dept_id) referencing t1initially de
impdp compress, table logging , performance test.
logging mode of table has effective in performance of impdp utility.test scenario:table test has rows 480W .1, expdp scott/tiger directory=diskc dumpfile=outp1.dmp tables=test --470MB2, expdp scott/tiger directory=diskc dumpfile=outp2.dmp tables=test compression=all--160MB, use compression to reduce I/O with same or less time.1, table with logging .impdp scott/tiger directory=diskc dumpfile=outp1.dmp tables=test time costs: 108s, and there are many archived log produced. 2, table with nologging
add hash partition , default tablespace for patitioned table
-- create a hash partition table.-- And specify a default tablespace for partitioned table.create table obj_pt tablespace users -- see follwing comment <1>partition by (object_id)(partition p1 tablespace ts_1, partition p1 tablespace ts_2 ); -- add hash partition alter table obj_pt add partition p3 ;alter table obj_pt add partition p4 tablespace ts_1;-- something relation to transport tablespace.<1> if you did not specify default tablespace for a partition table. Then it's rely on d
free list ,hwm
Oracle Freelist和HWM原理及性能优化http://www.knowsky.com/385108.html[@more@]
ROWDEPENDENCIES
ROWDEPENDENCIES:http://www.itpub.net/thread-982860-1-1.html# ORA_ROWSCNSQL> select rowid,ora_rowscn,a from t1;scn_to_timestamp(ora_rowscn) chg_time create table t_emp2 rowdependencies as select * from t_emp ;[@more@]
Oracle数据压缩
Oracle数据压缩http://blog.csdn.net/nhy520/archive/2008/11/05/3226519.aspx[@more@]
shrink space报ora-10635
shrink space报ora-10635http://www.linuxgoo.com/html/database/2009-05/561.html表test为分区表,只有一分区(pmaxvalue)。表处于users表空间内,而分区pmaxvalue位于users2内.表空间users和users2均为ASSM。alter table test enable row movement后,执行alter table test shrink space却报ORA-10635: invalid segment or tablespace type为何会这样呢?要来表是compressed ,我把表改为nocompressed 就可以了shrink_clause解释 shrink_clause 允许手动地释放表、索引组织表、索引、分区、物化视图和物化视图日志的空间。这个语句只对自
alter table move跟shrink space的区别
alter table move跟shrink space的区别http://blog.csdn.net/wangdeng1314/archive/2010/07/15/5735768.aspx[@more@]