加载实用程序vs reorg [英] Load utility vs reorg

问题描述

我们有一个更新的sql运行好几个月，在2分钟内完成

。昨天发生了变化，我们不得不在超过6小时后杀死这个

更新。更新输入表包含

正常100K行和主表包含大约135 M行。

我们重新排列主表，指定我们希望DB2在

更新，运行runstats并以相同的结果重新更新（有

取消它）。在reorg之后解释没有显示这个索引的使用。

我们然后重新加载主表并在10以下$ / b $ b秒内重新更新更新。解释显示了索引的使用。

我们认为reorg会解决问题，但它没有。

什么是负载做不同的事情？我们没有定义聚集索引。

谢谢，

Yuri

We have an update sql that was running OK for several months completing
in under 2 minutes . Yesterday something changed and we had to kill this
update after it ran more than 6 hours. Update input table contained the
normal 100K rows and master table contained about 135 M rows.
We reorged the master table specifying the index we wanted DB2 to use in
the update, ran runstats and reran the update with the same results (had
to cancel it). Explain after reorg did not show the use of this index.
We then reloaded the master table and reran the update in under 10
seconds. Explain showed the use of the index.
We thought that reorg would fix the problem, but it didn''t.
What is load doing differently? We don''t have clustered index defined.
Thanks,
Yuri

推荐答案

"由里" <宇** @ prodigy.net>在消息中写道

news：jG ***************** @ newssvr33.news.prodigy.co m ...

"Yuri" <yu**@prodigy.net> wrote in message
news:jG*****************@newssvr33.news.prodigy.co m...

我们有一个更新的sql运行好几个月完成
2分钟。昨天发生了变化，我们不得不在超过6小时后终止这个更新。更新输入表包含正常的100K行和主表包含大约135 M行。
我们重新排列主表，指定我们希望DB2在更新中使用的索引，运行runstats和reran具有相同结果的更新（已取消它）。在reorg之后解释没有显示该索引的使用。
然后我们重新加载主表并在10秒内重新进行更新。解释显示了索引的使用。
我们认为reorg可以解决问题，但它没有。
什么是负载做不同的？我们没有定义聚集索引。
谢谢，
Yuri

We have an update sql that was running OK for several months completing in
under 2 minutes . Yesterday something changed and we had to kill this
update after it ran more than 6 hours. Update input table contained the
normal 100K rows and master table contained about 135 M rows.
We reorged the master table specifying the index we wanted DB2 to use in
the update, ran runstats and reran the update with the same results (had
to cancel it). Explain after reorg did not show the use of this index.
We then reloaded the master table and reran the update in under 10
seconds. Explain showed the use of the index.
We thought that reorg would fix the problem, but it didn''t.
What is load doing differently? We don''t have clustered index defined.
Thanks,
Yuri

需要查看你的reorg语句以及你是否有一个聚类

索引在表格中定义。

Need to see your reorg statement and also whether you have a clustering
index defined on the table.

Mark A写道：

Mark A wrote:

" Yuri" <宇** @ prodigy.net>在消息中写道
新闻：jG ***************** @newssvr33.news.prodigy.co m ...

"Yuri" <yu**@prodigy.net> wrote in message
news:jG*****************@newssvr33.news.prodigy.co m...

我们有一个更新的sql运行好几个月完成
2分钟。昨天发生了变化，我们不得不在超过6小时后终止这个更新。更新输入表包含正常的100K行和主表包含大约135 M行。
我们重新排列主表，指定我们希望DB2在更新中使用的索引，运行runstats和reran具有相同结果的更新（已取消它）。在reorg之后解释没有显示该索引的使用。
然后我们重新加载主表并在10秒内重新进行更新。解释显示了索引的使用。
我们认为reorg可以解决问题，但它没有。
什么是负载做不同的？我们没有定义聚集索引。
谢谢，
Yuri

We have an update sql that was running OK for several months completing in
under 2 minutes . Yesterday something changed and we had to kill this
update after it ran more than 6 hours. Update input table contained the
normal 100K rows and master table contained about 135 M rows.
We reorged the master table specifying the index we wanted DB2 to use in
the update, ran runstats and reran the update with the same results (had
to cancel it). Explain after reorg did not show the use of this index.
We then reloaded the master table and reran the update in under 10
seconds. Explain showed the use of the index.
We thought that reorg would fix the problem, but it didn''t.
What is load doing differently? We don''t have clustered index defined.
Thanks,
Yuri

需要查看你的reorg语句以及你是否有一个聚类
在桌子上定义的索引。

Need to see your reorg statement and also whether you have a clustering
index defined on the table.

Mark，

我们没有定义聚类索引。在reorg语句中，我们指定了

索引，我们希望这个表被重新排列。

Mark,
we don''t have clustering index defined. In reorg statement we specified
index that we wanted this table to be reorged by.

" Yuri" <宇** @ prodigy.net>在消息中写道

新闻：Ol **************** @ newssvr33.news.prodigy.com ...

"Yuri" <yu**@prodigy.net> wrote in message
news:Ol****************@newssvr33.news.prodigy.com ...

Mark ，
我们没有定义聚类索引。在reorg语句中，我们指定了
索引，我们希望这个表被重新排列。

Mark,
we don''t have clustering index defined. In reorg statement we specified
index that we wanted this table to be reorged by.

您是否重新编写了索引？请发布您使用的重组声明。

Did you reorg the index? Please post the reorg statement you used.

这篇关于加载实用程序vs reorg的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！