如何处理行迁移？

方法一：
exp   ... imp
expdp ....impdp
--->影响应用，速度慢，安全

方法二：
move
--->简单，但是需要2倍磁盘空间来处理

方法三：

建立一个存放行迁移的表
SQL> @?/rdbms/admin/utlchain.sql

Table created.
SQL> desc chained_rows
 Name                                      Null?    Type
 ----------------------------------------- -------- ----------------------------
 OWNER_NAME                                         VARCHAR2(30)
 TABLE_NAME                                         VARCHAR2(30)
 CLUSTER_NAME                                       VARCHAR2(30)
 PARTITION_NAME                                     VARCHAR2(30)
 SUBPARTITION_NAME                                  VARCHAR2(30)
 HEAD_ROWID                                         ROWID
 ANALYZE_TIMESTAMP                                  DATE


SQL> analyze table scott.dept list chained rows;

Table analyzed.

建立中间表暂时存放迁移的数据
SQL> create table dept_bak as
  2  select * from scott.dept where rowid in
  3  (select HEAD_ROWID from chained_rows);

Table created.

删除原来表中所有的行迁移的行：
SQL> delete from scott.dept where rowid in  (select HEAD_ROWID from chained_rows);


再将被删除的行从中间表中反插回去
insert into scott.dept select * from dept_bak;


commit;