ETL的本质

innovate511 · 发表于 2006-1-8 23:58

不好意思，我对SQL SERVER的数据仓库产品不熟，虽然很多中小企业的OLTP系统使用SQL SERVER，不过数据仓库项目使用SQL SERVER数据仓库的，目前还不是很多，所以至今没接触过。感觉MS一贯的作风是操作方便，虽然性能一般，但是对于中小型项目还是有一定吸引力的。
我看了下介绍，感觉它更像MS自己提供的开发和管理工具，而不是项目设计，具体逻辑上的数据仓库和集市，还得做项目的人自己设计。

stillthere · 发表于 2006-1-9 04:00

非常感谢 innovate511君。

请问 innovate511君是否接触过缓慢变化维(The Slowly Changing Dimension transformation)？

很明显， MS 在cath up, 但愿它的普及商务智能(BI FOR MASSES)成功！

innovate511 · 发表于 2006-1-9 10:16

没有接触过，也许MS自己也搞出一些概念来，方便人们开发。这种普及有可能，MS可以在产品上进行整合，符合一些中小企业的小投资，相对高的回报的想法。

stillthere · 发表于 2006-1-9 10:34

希望有所帮助！

stillthere · 发表于 2006-1-9 10:51

The Slowly Changing Dimension transformation was covered in Kimball et al's book, Data warehouse toolkit...

see all the book titles from his website at

http://www.kimballuniversity.com/html/books.html

SSIS 2005 now deals with it, Cognos DecisionStream and its PowerPlay Transformer also deals with it.

The Slowly Changing Dimension transformation is an important concept many businesses run into and the tools have to deal with it.

But in my 6 years of DW/BI technical supporting work, I only run into it once or twice, so no wonder you did not see it before.

Anyhow, thank you very much, innovate511, for all your feedback. It was a great pleasure to read all your entries at itpub.net... I'd like to present you some roses;-)

26 or more years ago there was a song titled Dalian Hao that was composed by Wang Liping (大连好 by 王立平). I wonder if nowadays people are still singing it?

innovate511 · 发表于 2006-1-9 11:39

OK, I did not read <Data warehouse toolkit> yet, so many techniques and theories of DW/BI to learn.
I think you began to support DW/BI in 2000, right?

stillthere2 · 发表于 2006-1-9 23:44

元数据是纲。纲举目张。

举例：COGNOS的 DecisionStream, Impromptu Catalog 和COGNOS BI ARCHITECT以及ReportNet FRAMEWORK MANAGER就是用来管理元数据的。

纲举目张：运用元数据建立了后端DW的数据模型，前端的BI应用(报表，建立数据魔方，数据挖掘)就几乎是水到渠成。