Tags » Data Warehouse

Data vaults with Wherescape

           

   

Data Vaults have been gaining huge attention in recent years all over the world. Invented by Dan Linstedt, Data Vaults is truly an optimal way to build enterprise data warehouses.  23 more words

Data Warehouse

My Hot Tub Time Machine Summer

I’m one of those people who regularly keeps in contact with colleagues from the past…and by “past” I mean going way back to elementary school days in the 1960s and early 1970s growing up in Pittsburgh, all the way through professional colleagues and clients from only a few years ago. 964 more words

Analytics

4 Ways to Defeat Dogmatism in Your Company

Unless you’re Donald Trump—a fellow who’s always right, even when he’s wrong—there’s not much benefit to dogmatism. That’s because some of the biggest breakthroughs come as a result of challenging assumptions; especially those that are commonly accepted. 638 more words

Data Warehouse

工作日志 08252015

这就是我的工作日志开始的地方 :)

犹豫了很久还是写中文吧,每天没有太多时间留给过去,提高效率。日志中保护了诸多内容,不会出现任何产品、内部应用以及内部概念,同事、很多数据column都是用的化名。

今天重新加载了之前给未来项目做的一个transactional的fact table的prototype。三个月前写的代码和逻辑(transformation process)如今不work了,很可能是数据源的日期格式发生了变化。日期总是个棘手的问题。对于这个fact table,日期的逻辑有以下特殊之处:

Granularity是每月每条记录。从“点日期”变成“线”日期。
比如起始日期是06032014,被数据源标识为二季度(‘S2’),那么这条数据的时间跨度就会被定义为03012014-05012014。在前端三月到五月都会出现该条数据。就是说如果数据源是这样的:

vendor_id        start_date        quarter
01                      06032014           S2

ETL的output应该是这样的:

Date        vendor_id
03012014         01
04012014         01
05012014         01

这样处理的需求来源是:客户需要vendor信息按月显示。这样他们选择到这三个月的任何一个月的期间,该vendor的信息都会显示出来。其中,数据源日期不一定与最后定义的Date相符。
这样一条变多条,会让经过transformation的数据非常多,给load的server造成了很大负担,不过。。。需求嘛

果然一个小时是记不下来所有了。每天抽一点讲深入,日积月累吧。

周末会考虑出topic。

Work Diary

Do You Need a Single Version of the Truth?

A much touted benefit of business intelligence software is the ability to function as the system of record, to be the software that delivers the numbers that nobody can contest. 655 more words

Business Intelligence

Designing a Data Warehouse from the Ground Up Webinar Recording with Q & A

Thank you to everyone that registered and attended my webinar Designing your Data Warehouse from the Ground Up webinar this past Tuesday. And I’d also like to give a special thanks to my good friend, Mitchell Pearson ( 492 more words

SSAS

Who Should Lead Your Enterprise Big Data Program? (Part 4 of 4)

We come to the final entry of our four-part “mini-series” that discusses the characteristics someone leading your enterprise big data program should possess. Today: numbers 16-20. 994 more words

Analytics