前阵子遇到一个案例，需要将数据库中的几个表从USER A 移动到USER B下面，在ORACLE中，这个叫做更改表的所有者或者修改表的Schema。其实遇到这种案例，有好几种解决方法。下面我们通过实验来测试、验证一下。首先准备简单测试数据,如下所示:
SQL> CREATE TABLE TEST.KKK
( ID INT ,
NAME VARCHAR2(12) ,
CONSTRAINT PK_KKK PRIMARY KEY(ID)
);
Table created.
SQL> INSERT INTO TEST.KKK
2 VALUES(1000, 'kerry');
1 row created.
SQL> commit;
Commit complete.
方法1： 常规方法，在目标用户下创建表，并拷贝数据过去。
SQL> CREATE TABLE TEST1.KKK( ID INT, NAME VARCHAR2(12) ,CONSTRAINT PK_KKK PRIMARY KEY (ID));
Table created.
SQL> INSERT INTO TEST1.KKK
2 SELECT * FROM TEST.KKK;
1 row created.
SQL> COMMIT;
Commit complete.
SQL> DROP TABLE TEST.KKK;
Table dropped.
SQL>
当然也可以使用CREATE TABLE TEST1.KKK AS SELECT * FROM TEST.KKK; 但是使用这种方式需要注意，索引和约束都无法Copy过去。所以一定要慎用CREATE TABLE AS （CTAS）这种语法。
方法2： 使用expdp/impdp，导出表然后导入表修改表的Schema。使用exp/imp方式也是差不多，在此不做介绍。
[oracle@DB-Server dpdump]$ expdp system/xxxx tables=test.kkk directory=data_pump_dir dumpfile=kkk.dmp logfile=kkk.log;
Export: Release 11.2.0.1.0 - Production on Mon Jul 10 15:57:22 2017
Copyright (c) 1982, 2009, Oracle and/or its affiliates. All rights reserved.
Connected to: Oracle Database 11g Enterprise Edition Release 11.2.0.1.0 - 64bit Production
With the Partitioning, OLAP, Data Mining and Real Application Testing options
Starting "SYSTEM"."SYS_EXPORT_TABLE_01": system/******** tables=test.kkk directory=data_pump_dir dumpfile=kkk.dmp logfile=kkk.log
Estimate in progress using BLOCKS method...
Processing object type TABLE_EXPORT/TABLE/TABLE_DATA
Total estimation using BLOCKS method: 64 KB
Processing object type TABLE_EXPORT/TABLE/TABLE
Processing object type TABLE_EXPORT/TABLE/INDEX/INDEX
Processing object type TABLE_EXPORT/TABLE/CONSTRAINT/CONSTRAINT
. . exported "TEST"."KKK" 5.421 KB 1 rows
Master table "SYSTEM"."SYS_EXPORT_TABLE_01" successfully loaded/unloaded
******************************************************************************
Dump file set for SYSTEM.SYS_EXPORT_TABLE_01 is:
/u01/app/oracle/admin/gsp/dpdump/kkk.dmp
Job "SYSTEM"."SYS_EXPORT_TABLE_01" successfully completed at 15:57:26
[oracle@DB-Server dpdump]$ impdp system/xxxx tables=test.kkk directory=data_pump_dir remap_schema=test:test1 dumpfile=kkk.dmp logfile=import.log
Import: Release 11.2.0.1.0 - Production on Mon Jul 10 15:58:07 2017
Copyright (c) 1982, 2009, Oracle and/or its affiliates. All rights reserved.
Connected to: Oracle Database 11g Enterprise Edition Release 11.2.0.1.0 - 64bit Production
With the Partitioning, OLAP, Data Mining and Real Application Testing options
Master table "SYSTEM"."SYS_IMPORT_TABLE_01" successfully loaded/unloaded
Starting "SYSTEM"."SYS_IMPORT_TABLE_01": system/******** tables=test.kkk directory=data_pump_dir remap_schema=test:test1 dumpfile=kkk.dmp logfile=import.log
Processing object type TABLE_EXPORT/TABLE/TABLE
Processing object type TABLE_EXPORT/TABLE/TABLE_DATA
. . imported "TEST1"."KKK" 5.421 KB 1 rows
Processing object type TABLE_EXPORT/TABLE/INDEX/INDEX
Processing object type TABLE_EXPORT/TABLE/CONSTRAINT/CONSTRAINT
Job "SYSTEM"."SYS_IMPORT_TABLE_01" successfully completed at 15:58:09
注意上面方法无法移动表到其它表空间，所以，如果你也必须移动表到对应的表空间，那么就必须使用参数remap_tablespace， 如下所示：
impdp system/xxxxx tables=test.kkk directory=data_pump_dir remap_schema=test:test1 remap_tablespace=tbs_test_data:tbs_test1_data dumpfile=kkk.dmp logfile=import.log
方法3：修改系统表obj$、sys.con$等 （这个仅仅作为实验测试而已，不可应用于生产环境）,ASK TOM里面强烈不建议使用这种方法，原文如下，不过不妨碍我们在测试环境玩一玩，了解一下。
SQL> select obj#, owner#, name, namespace from obj$ where name='KKK';
OBJ# OWNER# NAME NAMESPACE
---------- ---------- ------------------------------ ----------
93220 85 KKK 1
SQL> select user_id , username from dba_users where username in ('TEST', 'TEST1');
USER_ID USERNAME
---------- ------------------------------
86 TEST1
85 TEST
SQL> update obj$ set owner#=86 where obj#=93220;
1 row updated.
SQL> commit;
Commit complete.
SQL>
如下所示，更新了系统表obj$后，你会发现TEST.KKK与TEST1.KKK两个表都存在，这个时候可以执行ALTER SYSTEM FLUSH SHARED_POOL命令后，TEST.KKK就不存在了。
SQL> select * from v$mystat where rownum=1;
SID STATISTIC# VALUE
---------- ---------- ----------
12 0 0
SQL> SELECT * FROM TEST.KKK;
ID NAME
---------- ------------
1000 kerry
SQL> SELECT * FROM TEST1.KKK;
ID NAME
---------- ------------
1000 kerry
SQL> ALTER SYSTEM FLUSH SHARED_POOL;
System altered.
SQL> SELECT * FROM TEST.KKK;
SELECT * FROM TEST.KKK
*
ERROR at line 1:
ORA-00942: table or view does not exist
SQL> SELECT * FROM TEST1.KKK;
ID NAME
---------- ------------
1000 kerry
SQL>
另外，更新了系统表obj$后，你会发现索引、约束信息还没有变化，如下所示，OWNER依然为TEST，索引也可以在obj$中更新对应记录，但是约束就必须继续查找
SQL> SELECT OWNER ,INDEX_NAME , TABLE_NAME FROM DBA_INDEXES WHERE TABLE_NAME='KKK';
OWNER INDEX_NAME TABLE_NAME
------------------------------ ------------------------------ ------------------------------
TEST PK_KKK KKK
SQL> COL OWNER FOR A20;
SQL> SELECT OWNER, CONSTRAINT_NAME FROM DBA_CONSTRAINTS WHERE TABLE_NAME='KKK';
OWNER CONSTRAINT_NAME
-------------------- ------------------------------
TEST PK_KKK
SQL> select obj#, owner#, name, namespace from obj$ where name='PK_KKK';
OBJ# OWNER# NAME NAMESPACE
---------- ---------- ------------------------------ ----------
93221 85 PK_KKK 4
SQL> update obj$ set owner#=86 where obj#=93221;
1 row updated.
SQL> commit;
Commit complete.
SQL> SELECT OWNER ,INDEX_NAME , TABLE_NAME FROM DBA_INDEXES WHERE TABLE_NAME='KKK';
OWNER INDEX_NAME TABLE_NAME
-------------------- ------------------------------ ------------------------------
TEST1 PK_KKK KKK
SQL> SELECT OWNER, CONSTRAINT_NAME FROM DBA_CONSTRAINTS WHERE TABLE_NAME='KKK';
OWNER CONSTRAINT_NAME
-------------------- ------------------------------
TEST PK_KKK
SQL>
而DBA_CONSTRAINTS的定义位于 cdcore.sql中（$ORACLE_HOME/rdbms/admin)中，具体定义如下所示所示（本文测试环境为Oracle 11g），从其定义可以知道，我们必须更新sys.con$中的记录。
create or replace view DBA_CONSTRAINTS
(OWNER, CONSTRAINT_NAME, CONSTRAINT_TYPE,
TABLE_NAME, SEARCH_CONDITION, R_OWNER,
R_CONSTRAINT_NAME, DELETE_RULE, STATUS,
DEFERRABLE, DEFERRED, VALIDATED, GENERATED,
BAD, RELY, LAST_CHANGE, INDEX_OWNER, INDEX_NAME,
INVALID, VIEW_RELATED)
as
select ou.name, oc.name,
decode(c.type#, 1, 'C', 2, 'P', 3, 'U',
4, 'R', 5, 'V', 6, 'O', 7,'C', 8, 'H', 9, 'F',
10, 'F', 11, 'F', 13, 'F', '?'),
o.name, c.condition, ru.name, rc.name,
decode(c.type#, 4,
decode(c.refact, 1, 'CASCADE', 2, 'SET NULL', 'NO ACTION'),
NULL),
decode(c.type#, 5, 'ENABLED',
decode(c.enabled, NULL, 'DISABLED', 'ENABLED')),
decode(bitand(c.defer, 1), 1, 'DEFERRABLE', 'NOT DEFERRABLE'),
decode(bitand(c.defer, 2), 2, 'DEFERRED', 'IMMEDIATE'),
decode(bitand(c.defer, 4), 4, 'VALIDATED', 'NOT VALIDATED'),
decode(bitand(c.defer, 8), 8, 'GENERATED NAME', 'USER NAME'),
decode(bitand(c.defer,16),16, 'BAD', null),
decode(bitand(c.defer,32),32, 'RELY', null),
c.mtime,
decode(c.type#, 2, ui.name, 3, ui.name, null),
decode(c.type#, 2, oi.name, 3, oi.name, null),
decode(bitand(c.defer, 256), 256,
decode(c.type#, 4,
case when (bitand(c.defer, 128) = 128
or o.status in (3, 5)
or ro.status in (3, 5)) then 'INVALID'
else null end,
case when (bitand(c.defer, 128) = 128
or o.status in (3, 5)) then 'INVALID'
else null end
),
null),
decode(bitand(c.defer, 256), 256, 'DEPEND ON VIEW', null)
from sys.con$ oc, sys.con$ rc, sys."_BASE_USER" ou, sys."_BASE_USER" ru,
sys."_CURRENT_EDITION_OBJ" ro, sys."_CURRENT_EDITION_OBJ" o, sys.cdef$ c,
sys.obj$ oi, sys.user$ ui
where oc.owner# = ou.user#
and oc.con# = c.con#
and c.obj# = o.obj#
and c.type# != 8 /* don't include hash expressions */
and (c.type# < 14 or c.type# > 17) /* don't include supplog cons */
and (c.type# != 12) /* don't include log group cons */
and c.rcon# = rc.con#(+)
and c.enabled = oi.obj#(+)
and oi.owner# = ui.user#(+)
and rc.owner# = ru.user#(+)
and c.robj# = ro.obj#(+)
/
那么我们必须更新sys.con$中记录，如下所示：
SQL> select owner#,name, con# from sys.con$ where name='PK_KKK';
OWNER# NAME CON#
---------- ------------------------------ ----------
85 PK_KKK 19023
SQL> update sys.con$ set owner#=86 where name='PK_KKK';
1 row updated.
SQL> commit;
Commit complete.
SQL> COL OWNER FOR A32;
SQL> SELECT OWNER, CONSTRAINT_NAME FROM DBA_CONSTRAINTS WHERE TABLE_NAME='KKK';
OWNER CONSTRAINT_NAME
-------------------------------- ------------------------------
TEST1 PK_KKK
SQL>
OK，到此基本的几种方法我们已经介绍了,实际场景当中，那些需要修改Schema的表可能会非常大，生产环境完全不能使用方法3这种捷径。如果使用CTAS这种方法，即使使用NOLOGGING和开启并行，也是非常消耗资源和时间的。使用expdp/impdp 稍微好一些，但是也不能避免一些资源开销。 还有一种方法就是使用exchange partition，exchange只是在ORACLE的数据字典中修改了分区和表的结构。数据并未发生任何修改，因此速度很快.对于大表来说，非常高效！如下测试案例所示：
SQL> create table test.big_table
2 as
3 select object_id, object_name from dba_objects;
Table created.
SQL>
SQL>
SQL> create table test1.big_table
2 ( object_id number,
3 object_name varchar2(128)
4 )
5 partition by range( object_id)
6 (
7 partition big_table_par values less than(maxvalue)
8 );
Table created.
SQL> alter table test1.big_table exchange partition big_table_par
2 with table test.big_table;
Table altered.
SQL> select count(*) from test1.big_table;
COUNT(*)
----------
72329
SQL> select count(*) from test.big_table;
COUNT(*)
----------
0
SQL>
如上所示，我们假设test.big_table是一个非常大的表，那么使用exchange partition就能快速、高效地将一个大表修改Schema,唯一不足的是，需要将普通表修改为分区表，当然，很多时候，大表很可能已经是分区表，使用这种方式不会有任何修改。另外，我们上面都是测试普通表，其实如果是分区表，使用exchange partition则是最佳选择！
参考资料：
https://asktom.oracle.com/pls/apex/f?p=100:11:0::::P11_QUESTION_ID:752030266230