A cookbook builds upon reader's knowledge on some subject to teach her new recipes. For instance, have you ever seen some TV cook teaching how to break eggs? You just learn it from seeing it and then it's recipe time!

Packt new book, Pentaho Data Integration Cookbook 2nd Edition (http://www.packtpub.com/pentaho-data...d-edition/book which they presented me for reviewing), released a couple of weeks ago is the quintessential cookbook: it will not teach you how to install, run or create new transformations and jobs - that is a job for the great PDI Beginner's Guide (also from Packt.) Instead it is going to teach you recipes, little HOW-TOs, on over a hundred subjects.

There are recipes about connecting to a database, how to parametrize it (and why do it), how to read data from relational or NoSQL (like MongoDB) databases, how to feed it to a sub-transformation, then handling the streams and its metadata up to finally writing everything to a Hadoop cluster and so on. It does not stop on PDI: There are recipes on how to read and write data to cloud (Amazon and Salesforce), how to run jobs and transformations from inside BI Server (effectivelly turning it into a data integration resource), generate fully formated Pentaho Reports from inside a PDI transformation OR with a transformation as a data source from inside PRD.

I've been reading the book for the last two weeks and, frankly, it seems endless: Data analysis, optimization, sampling data for Data Mining, InstaView, AgileBI, read SAS files, treating JSON and XML files, creating new functionalities (requires Java understanding but NO compilation or program building), send e-mail, add extra log messages etc. etc. etc. Besides the sheer recipe number, each recipe comes with some variation on the theme, occasionally a discussion about the pros and cons of the possibile options and alternatives (like Database Lookup vs. Database Join) So, each recipe not only tells you how to do something, but also adds a lot of value explaining how PDI works. When applicable there are comments on performance issues and how to get the best out of PDI (my favorite one is on location 400: how to use a in-memory database to speed up lookups!)

You can take a look at the complete recipe list on Packt's site (at http://www.packtpub.com/pentaho-data...d-edition/book).

Every recipe builds on the reader's knowledge but there is enough details so no novice user can run the recipe, without boring the more experienced reader. Figures are sparingly used, only when the authors feel it is going to better explain some setup. It helps to keep the book shorter (a +400-page hulk already) but on the other hand demands more attention on reading.

As a lot of Packt books, this one has been written by non-native English speakers which renders the language a bit "unfluid". It does not hinder the reading but makes it unconfortable some times.

This book (here http://www.packtpub.com/pentaho-data...d-edition/book) is a treasure chest, with great value to either new or seasoned PDI users. If you already know how to use PDI and want to get better with it, this book is a good reading. If you want to learn how to use PDI, look for Packt's "PDI Begginer’s Guide" and then read this book. It is a must-have!