Customer Reviews

Joe Celko's Trees and Hierarchies in SQL for Smarties, (The Morgan Kaufmann Series in Data Management Systems)

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

 

 

I love and savor each of Celko?s books for their thoroughness, depth and surprises. However, with this book I am little disappointed because it showed his favorite solutions and omitted many common solutions that are better in various real-world situations. Most algorithms provided in the book are for overnight processing, not real-time update.

Omitted is the most common way hierarchies are represented in Data Warehouses using a "hierarchy bridge table". See Kimball?s book "Data Warehouse Toolkit" for more detail (yet not enough detail to give a Celko-like exploration of the topic). The bridge table solution trades away storage space for greater speed by creating a record for every path enumeration.

The book?s primary focus is on strict hierarchies. Not enough attention is given to convergent graphs and other arbitrary directed acyclic graphs, like bill of materials that reuse assemblies, where the nested sets model fails (p.164).

Also missing is maintenance of historical versions of the hierarchy, often required by financial applications.

Section 9.3 on the extremely powerful DB2 ?WITH? operator is too slim, especially since it is a SQL-99 standard and is now available with Microsoft SQL Server. If you work with hierarchies or acyclic graphs in DB2 or SQL Server take the time to learn how to use ?common subquery expressions?.

By all means if you work with hierarchies you must buy this book. No doubt the Second revison of this book will blow us away.

When I started looking into managing hierarchical data I found a lot of short articles that provided bits and pieces of the big picture, but nothing gave a good in-depth review of everything I needed to know to handle hierarchies.

One trend that began to develop was that the best information on the subject was coming from Joe Celko, a rather prolific author/speaker on all things SQL.

Joe Celko has a book in publication called SQL For Smarties, and this book had some basic information on handling hierarchies, which he followed up with Trees and Hierarchies in SQL for Smarties, a truly excellent book on the subject.

This book is not for the faint of heart, but it really is a must-read for those who are looking to handle hierarchies in SQL.

The book only addresses MySQL once, and while in that section it repeats Joe's opinion of MySQL not being a real database, users of MySQL 5 can rest assured that the stored procedure examples in the book work without modification (thanks to Joe and MySQL 5 both using the standard PL/PSM stored procedure syntax).

Overall this is an excellent book for those looking to implement hierarchical data in SQL.

While a little advanced (... the title does say for smarties...) this is one of the few IT books I have ever bought which I read more than once. It's a great mix of brain baking concepts and memory jarring discussions of things you probably slept through your sophomore year. If you've been working with or teaching yourself SQL and are looking to move on to the next step, Celko's books are an excellent place to start.

Note to MySQL users - as of MySQL 5 you can create everything written in this book, although you will need to remove the ATOMIC keyword from the procedures.

The implementations for trees in SQL described in this book are great.

However if you don't need to understand it really deep you just need to read Celko's article that you can easily find around the web.

I bought this book awhile back and skimmed through it. The last month I have been trying to get a clients adjacency list database to run fast enough for reporting and used the nested set. I has been a great solution to this problem.
While I could have found it on the web, I wouldn't have known what I was looking for with out reading about it first.

It's not what I expected.

I'm doing alot of SQL programming now and there's also alot of self-joins and just simply alot of analytics in my new position using SQL. I thought this book would help me in this area but was wrong.

If you have computer science background think of this book as trying to do binary trees and linked lists in one table. It uses one table fields to try to get to the left and right nodes and back to the root node. Don't get me wrong, if this is what you need to do then this book is right for you. However, if you want to do real life advanced application programming using SQL to get your job done I would highly recommend to review other books. Sincerely, JP