PostgreSQL Development Essentials: Advanced querying, data modeling and performance tuning

While I'm not going to completely disagree with PK's review, I'm also going to point out that it doesn't know the story of this book. I'm one of the technical reviewers listed in the foreword of the book. There are three things one needs to know about this specific book. First, there are two authors instead of one because the first author just disappeared from Earth after writing a few chapters from the outline. I was asked to complete the book, but for reasons unknown (I suspect "cost of labor"), someone else did it instead. Fine, I'm not holding a grudge over that at all; anyone can self-publish if they want. Second, the critique of grammar in the book is fair, but oh man if you had seen the first drafts... The grammatical editors at Pakt are truly amazing. Unfortunately, they sometimes also change the meaning of the sentences while they're fixing the original grammar. Seriously, the grammatical editors rewrote an easy 80% of this book. And that brings me to the third point - the process. The way this works is: the author writes up a draft, the draft goes to the reviewers, the reviewers add comments, the author integrates those comments at their discretion, then that integrated result goes to the publisher's editors to clean up and arrange. So the reviewers listed in the front of the book, like me, can theoretically be amazing (not that I'm necessarily "amazing", but I do have sufficient competence to manage these simple examples). That doesn't mean their input is necessarily getting to the published book unless 1) the author integrates the feedback *and* 2) the editors later don't mangle things to make the examples appear incorrect.I can assure anyone who's reading this that the examples *all* were verified as working properly and fit the way I interpreted the author when I handed my notes back. Don't send all of the names from the book down with the ship; some of us only got paid in the form of a free book for our time (which was significant on this one), and had no control of the final content. :D

This book consists of a random selection of semi-literate re-hashes of the user manual and Wikipedia pages, along with a few code examples, most of which are broken.The authors jump from one topic to the next, never explaining anything in any real detail. For example, chapter one flits aimlessly from creating views, to creating cursors, to the GROUP BY clause to ‘complex’ topics such as subqueries and joins. This must be the only book on any database in which cursors are mentioned in the first few pages.Here are a few examples of how wretched the text is. For example, the section on cursors says:“It then uses a FETCH statement with the NEXT keyword to select the fifth row, and then another FETCHstatement with the PRIOR keyword to again select the fourth retrieved row.’To illustrate this, the authors provide the following:FETCH 4 FROM order_cur;Clearly, some code has been missed out, and nobody has noticed.In the section on JSON, several lines have again been missed out and/or replaced with the wrong code:“Let's create a normal table and insert data into it:CREATE TABLE product_info (id serial, product_name text, product_categorytext, price int);INSERT INTO product_info (product_name ,product_category , price)VALUES ('Chair', 'Furniture', 200);Now, since we have inserted the record in a normal way, we will change this to the JSONformat. We will use the row_to_json function:row_to_json-----------------------------------------------{"f1":"Chair","f2":"Furniture","f3":200}(1 row)”I have no idea if the authors actually believe that typing ‘row_to_json’ would produce the result they claim: it obviously doesn’t. Nor does it work like this inside the obviously needed select statement. The next few lines are even more bizarre:“Similarly, we can output the array values as JSON with the help of the array_to_json function:CREATE TABLE product_info(id serial, product_category text, producttext[]);INSERT INTO product_info (product_category , product) VALUES('Stationary', '{"Pen","Pencil"}');”Nothing is being output as JSON: instead, a new table is being created using JSON instead of text.Here’s another strange explanation:SELECT product, SUM(sale) AS "Total sales"FROM order_detailsGROUP BY product;‘In the select statement, we have sales where we applied the SUM function and the other fieldproduct is not part of SUM, we must use in the GROUP BY clause.’The authors are attempting to say “we must group on a field other than the one which we are aggregating: we use ‘group by’ on the product field and use the aggregate sum() function on the ‘sale’ field’”.Here’s yet another example where the explanation is incomprehensible:“The HAVING clause is used to specify which individual group(s) is to be displayed, or in simple language we use theHAVING clause in order to filter the groups on the basis of an aggregate function condition.”Compare this drivel with the clear explanation in the (free) user manual:“WHERE selects input rows before groups and aggregates are computed (thus, it controls which rows go into the aggregate computation), whereas HAVING selects group rows after groups and aggregates are computed.”The section on transactions fails to explain either how isolation levels work or which ones PostgreSQL implements. The explanation of serializable isolation is particularly inadequate:‘As you saw in the preceding table, moving from the Read Uncommitted to Serializable level, the phenomena behavior is reducing.’In those cases where there isn’t a Wikipedia page as a source, the text descends into meaningless waffle. For example:“In a largely built database, there are cases where performance may not be up to the desiredexpectations. And, achieving more practical goals may spoil the purity of the design. Thesteps for the optimized design should be considered with the utmost priority before youeven consider any optimization.”Clearly, neither of the authors has a proper grasp of English, and neither has enough experience of PostgreSQL to be able to write a useful book on the subject. Unbelievably, this garbage has been checked by over a dozen people. The book boasts a copy editor, two reviewers, a proof-reader, a technical editor, a content development editor, a production coordinator, a commissioning editor, an acquisition editor and a production coordinator. All of these people must be even more inept at their job than the authors or else asleep.