oracle分析函数删除重复数据

历史库有张分区表有二十亿数据，有300w数据是重复的，开发写的删除脚本效率比较低。用分析函数改写删除脚本，按分区来删除。

select *

  from TAB

where rowid in (select rid

                   from (select rowid as rid,

                                RCVBL_AMT_ID,

                                row_number() over(partition by RCVBL_AMT_ID order by rowid desc) as seq

                           from TAB t

                          where t.ORG_NO LIKE '32405' || '%'

                            AND t.RCVBL_YM = '201403')

                  where seq > 1);