Customer Reviews

The Data Warehouse ETL Toolkit: Practical Techniques for Extracting, Cleanin

5 star:		(13)
4 star:		(2)
3 star:		(0)
2 star:		(0)
1 star:		(0)

 

 

In this book Ralph lays down a framework for constructing the DW ETL. This is useful not just in constructing quality ETL processes, but also because Ralph's works tend to 'set' standards in data warehousing. The format of this book is similar to the Lifecycle Toolkit. Ralph takes a very staged, logical approach to the material. Some sections are just great e.g. the chapters on Extraction and Development. A small amount of the material is repeated from the Lifecycle Toolkit and Dimensional Modeling books, but no more than is needed to make this book stand on its own.

Also like the other books, this one takes a vendor agnostic approach. While this may increase the shelf-life of the book, I would have appreciated some comparisons between the major vendors out there today.

Overall: I recommend this one as a buy, even if you have Ralph's other books.

This is one of the few references out there providing the building blocks of good ETL design. There is plenty of technical documentation and forums out there that are specific to one ETL tool or DBMS but this is a better starting place for ETL developers. It is required reading as ETL projects often take short cuts in design, data quality and metadata management and reporting. This leads to very expensive Data Warehouse administration costs and often a complete rebuild of load jobs.

The book is relevent for people using most ETL or ELT tools and it will remain relevent for years even as the ETL products continue to advance and mature. It is targeted at DW but the basic flow of Extract, Clean, Conform and Deliver is suitable for most types of data loads.

Good coverage of the alternatives to traditional overnight bulk loads in the section on real-time ETL systems (also describes Microbatch) as the businesses and the major ETL vendors shift to SOA.

This book takes almost all issues in a data warehouse design and represents them oriented to ETL features. Actually, ETLing matches the whole of the data warehouse (more or less), so the need to describe them makes this book an autonomous work you can read without referring to previous books by Kimball. Besides, I think that some technical descriptions have been better performed here: in my experience it is impossible to undertake dwh activities without (at least) a sound knowledge about general features (indexes, use of a bulk loader vs. INSERT, etc.) of RDBMS, and this paper addresses them conveniently. On the other hand, the flat style used lacks to give evidence to the very significant issues, which happen so to be mixed up with less important statements; that demands to pay high attention while reading, but a blurring boundary between subtleties and trivialities seems to be a common shortcoming in dwh literature. Even with that flaw, the ETL Toolkit turn out as an outstanding reference to state of the art of dwh technology.

For a book which has "ETL Toolkit" in the name, I would think this would get right down to the nuts and bolts of how to write an ETL system. Yet, after reading this book, I still have the feeling that I still wouldn't know how to build one. From the book, I know what the issues are about inserting, deleting, and organizing the data, but how does that translate into SQL which actually does the work? There are hardly any SQL or sample code to go by. Issues like dealing with multiple database sources are mentioned, but just what do you do when cities are mentioned in both databases and they don't quite match up? If you decide to roll your own, the book really doesn't give you any basic framework to follow. This book does provide important insights on ETL systems, but don't expect to be able to write one based on this book. For example, you are told to bulk insert data, but you are also expected to exchange all of the natural keys with surrogate keys. If you have to examine each line and exchange keys, you cannot bulk load the data and this ends up being a very CPU intensive process where each row is laboriously examined, exchanged and then inserted one row at a time. This is a very practical ETL difficulty but yet, the book doesn't deal with this problem. It just assumes this won't be a problem for you. In some ways, the book is more about selling you the 'Kimbal' way of data warehouses as other more intuitive ways of constructing a data warehouse are not even considered. Like a sales pitch and a fancy powerpoint presentation, it lacks the substance to actually create a well performing ETL process. That is left as an exercise for the reader.

Ralph Kimball has rounded out his complete recipe for building fast, cost effective, robust and durable enterprise dimensional data warehouses with this immensely valuable addition to all IT & Data Warehouse professionals' bookshelves.

Without a doubt ETL has been the biggest stumbling block to deploying and maintaining well architected data warehouses that stand the test of time. Ralph draws on his years of experience and engagement with thousands of projects and crystallizes the `Best Practices' into an effective application architecture for all ETL systems regardless of what tools projects use for implementation.

In this thorough examination of the Extract, Transform and Load (ETL) process, Ralph identifies 38 critical functions that all ETL systems need to implement for success in the long haul. He thoughtfully lays out simple and practical approaches for how each of these functions can be implemented by projects with any size of budget.

For many, the paradoxical nature of ETL (seeming trivial yet replete with endlessly complex details that constantly change) has been the proverbial straw that has broken the bank for many DW projects. Continual customer pressure to grow, improve performance, and quickly deal with changing business conditions have left developers and architects grasping for more powerful and flexible approaches to ETL that meet project timelines, yet evolve and improve with age. Armed with this enlightening roadmap, many DW professionals will be far better equipped to design and build systems that meet the challenges today and tomorrow.

I am currently working as an ETL Developer at a company
Fourier Approach, Centurion, South Africa.

Most of the time this is a fairly hot seat -
because so many business requirements are dependant on the
Quality of Information produced by the ETL process.

I always asked myself,

* Am I doing the right thing?
* Is this the best solution?
* How would other developers do this?

A while ago I attended the course

"ETL Architecture and Design Workshop"

presented by Joe Caserta, and hosted by Alicornio Africa in Johannesburg, South Africa.
Before the presentation we received a copy of the book
"The Data Warehouse ETL Toolkit".

This changed my whole perspective.
The book adressed all my ETL questions,
with examples from real-world situations.
It covers the whole ETL process and gives answers
to almost every question you will ever think of asking.

I must say this is a very handy tool on the desk of any serious ETL Developer.

Regards,

André Ackermann
ETL Developer

This book gives practical guidelines to follow through the ETL cycle, it does not matter if you are using an Industry Standard ETL tool or writing your own ETL process from scratch, this book will be useful for both. I found it very useful. Definitely worth a read for anyone who is new to ETL.

I've been doing this for 9 years, and this is the best book I've seen on ETL procesing and its role in data warehousing. Before you start your data warehouse effort have your team read this book. Not just the ETL members...both the front/back room technicians will benefit from understanding the clear presentation of what can be in scope (dare I say, "best practice" yikes!) for ETL processing. The ideas for capturing highly valuable data quality and cleansing processes is no less helpful than the emphasis on loading the data unless it is misleading enmasse or so obviously flawed. The bias toward driving the data to the front room for presentation forces data quality issues to the surface where they must be dealt with and the loop to operational systems (or perhaps even flawed ETL transforms!) closed. Illustrates alternatives for dealing with the messy reality of suspended data, late arriving facts and dimensions. Use it as a guide for your ETL efforts. Buy this book, read it and then buy a few more for your team if you agree.

This is not only a great book about ETL, but one of the best books of the "Data Warehouse Toolkit" series.

I strongly recommend this book if you are somehow involved with the construction of a Data Warehouse.

Read this book BEFORE deploying your ETL tool. It explains the critical objectives and methods for designing ETL procedures in simple to follow language. I was surprised to see how much of the book is devoted to data quality assurance, but it made sense after reading it.