Sybase ASE数据清除批处理 - 设计和性能(Sybase ASE data purge batch - design & performance)

 我正在研究Sybase ASE（迁移到15.7）数据清除实用程序，供多个表/数据库使用，以删除大量不需要的旧数据。 
 
 
 
 收到输入表名后，自动找出子表并删除数据。 但是，我找不到像Oracle的“Connect by ... Prior”子句这样的分层查询子句。 有没有其他方法来实现这个？ 
 
 
 我通过以小增量循环遍历多个事务/提交来删除数据。 删除后，在什么时间间隔，我应该“重组重建”？ 
 
 
 我需要做更新统计吗？ 如果必须，在进行更新统计之前我应该​​考虑的标准是什么？ 
 
 
 某些表可能已分区。 在分区的角度来看，我需要考虑什么吗？ 
 
 
 我们的一些数据库（我猜索引......？）是聚集的。 我对集群没有太多了解。 我是否需要在聚类视角中考虑任何事情？ 
 
 
 在处理结束时发送电子邮件。 内置的电子邮件包是否类似于oracle的UTIL_SMTP？ 
 

I am working on a Sybase ASE (migrating to 15.7) data purge utility to be used by multiple tables/ databases to delete huge amount of unwanted older data.
 
 
 
After receiving an input table name, automatically figure out the child tables and delete data. But, I couldn't find an hierarchical query clause like Oracle's "Connect by .. Prior" clause. Is there any other way to implement this?
 
 
I am deleting data by looping through multiple transaction/ commits in small increments. After the deletes, at what interval, should I do "reorg rebuild"?
 
 
Do I need to do update statistics? If I have to, what is the criteria that I should consider before doing update statistics?
 
 
Some tables may be partitioned. Is there anything that I need to consider in partition's perspective?
 
 
Some of our DB's (i guess index..?) are clustered. I don't have much idea about clustering. Do I need to consider anything in clustering perspective?
 
 
Send Email at the end of processing. Does built-in email package similar to oracle's UTIL_SMTP?
 

原文：https://stackoverflow.com/questions/15692299