oracle数据泵、表空间、库导入导出

2014-04-28

１．查找文件

[oracle@oracle53 ~]$ which expdp         

/u01/app/oracle/11.2.0/db_1/bin/expdp

[oracle@oracle53 ~]$ locate expdp

/u01/app/oracle/11.2.0/db_1/bin/expdp

/u01/app/oracle/11.2.0/db_1/bin/expdpO

[oracle@oracle53 ~]$ whereis expdp

２．数据泵导出

##hr用户

shell>expdp hr/hr directory=DATA_PUMP_DIR dumpfile=cc.cc tables=hr.jobs

ORA-39087: directory name DATA_PUMP_DIR is invalid

##系统用户

shell>expdp system/322815 directory=DATA_PUMP_DIR dumpfile=cc.cc tables=hr.jobs

**********************************************************************

Dump file set for SYSTEM.SYS_EXPORT_TABLE_01 is:

  /u01/app/oracle/admin/orcl/dpdump/cc.cc

Job "SYSTEM"."SYS_EXPORT_TABLE_01" successfully completed at 09:09:02

##也能够不指定目录

shell>expdp system/322815  dumpfile=cc.cc tables=hr.jobs

##基于scott用户

shell>expdp scott/tiger directory=DATA_PUMP_DIR dumpfile=scott.dmp schemas=scott

３．建立目录对象　－－须要先建立/u01/expdp目录　

SQL> create directory exp_home as '/u01/expdp';

SQL> grant read,write on directory exp_home to scott;

SQL>select  * from DBA_DIRECTORIES;

４．全库导出和导入　　－－须要EXP_FULL_DTABASE，IMP_FULL_DTABASE角色

导出：expdp scott/tiger directory=exp_dir full=y dumpfile=full_db.dmp 

导入：impdp scott/tiger directory=exp_dir full=y dumpfile=full_db.dmp

５．表空间的导出和导入

导出:expdp scott/tiger directory=exp_home dumpfile=tbs_01.dmp tablespaces=tbs_01;

／／若是删除表空间后恢复表空间数据，须要先建立表空间再恢复表空间数据

SQL> drop tablespace tbs_01 including contents;　--不会物理删除磁盘上的datafile

##查看datafile并删除

shell>cd /u01/app/oracle/oradata/orcl/

shell>rm tbs_01.dbf 

##导入

[oracle@ora243 expdp]$ impdp scott/tiger directory=exp_home dumpfile=tbs_01.dmp tablespaces=tbs_01

错误显示：ORA-00959: tablespace 'TBS_01' does not exist

##建立表空间

SQL>create tablespace tbs_01 datafile '/u01/app/oracle/oradata/orcl/tbs_01.dbf' size 100m;

##导入

impdp scott/tiger directory=data_pump_dir dumpfile=tbs_01.dmp tablespaces=tbs_01

６．schema的导入和导出

$ expdp scott/tiger directory=exp_home dumpfile=hr.dmp schemas=hr

$ impdp scott/tiger directory=exp_home dumpfile=hr.dmp schemas=hr

７．表的导入和导出

expdp scott/tiger directory=exp_home dumpfile=tble_emp.dmp tables=scott.emp reuse_dumpfiles=y

impdp scott/tiger directory=exp_home dumpfile=tble_emp.dmp tables=emp

８．数据泵导出参数

reuse_dumpfiles:默认不覆盖转储文件，覆盖用resue_dumpfiles=y

compression:指定压缩那些数据，在导出数据写入转储文件前．data_only只压缩全部数据，metadata_only,只压缩全部元数据（表结构）

//扩展

##查看角色权限

SQL>select * from role_sys_privs where role='EXP_FULL_DTABASE';

SQL>select * from role_sys_privs where role='CONNECT'

ROLE       PRIVILEGE ADM

------------------------------ ---------------------------------------- ---

CONNECT       CREATE SESSION NO

metalink:oracle内部的付费帐号

９．导出hr.employees用户中JOB_ID=IT_PROG,薪水大于8000的雇员信息　　－－加query参数

shell>expdp hr/hr directory=data_pump_file dumpfile=emp.dmp tables=hr.employees query=hr.employees:"where job_id='IT_PROG' and salary>8000"

出错信息：

LRM-00101: unknown parameter name 'job_id'

正解：

expdp scott/tiger directory=data_pump_dir dumpfile=emp.dmp tables=hr.employees query=hr.employees:\"where JOB_ID\=\'IT_PROG\' and SALARY\>8000\"

１０．SCN系统更改号

数据泵：精确到行记录，只能还原到备份的状态

rman:最小粒度为文件，和归档日志结合，能够还原到任什么时候间点．

##查看当前SCN

sql>select current_scn from v$database;

CURRENT_SCN

-----------

    2280109

##插入数据

sql>insert into dept values(22,'ha','luoyang');

##查看更改过的SCN

sql>select current_scn from v$database;

CURRENT_SCN

-----------

    2280147