It is a course book: it starts with the concepts of the relational model, moves to simple queries and works up through operators, functions, grouping and the rest of the language, with exercises and self-assessment in every chapter. The edition in circulation is a scan with recognition errors. None of that matters for what it is good for.
What I took from it
Its treatment of null values, which gets a whole section with truth tables. In daily work, half the reports that do not add up are reports where someone compared against a value that does not exist and got silence instead of an error. Having that explained in Spanish, with examples, is what turns the conversation with whoever asked for the report into understanding rather than "the system is broken".
The second thing is about my own practice: this is what I recommend when someone in finance wants to stop requesting reports and start writing them. You do not need to teach them a database engine or a fashionable tool; you need to teach them to read a query and know what it returns. After that, meetings change tone.
And a third, less expected: the exercises are designed to be solved by hand. Writing a query without autocompletion forces you to know what things are called, and it shows later.
What I left out
Everything that is not standard: engine-specific functions change, and the book is well over a decade old. There is also nothing on performance, execution plans or large data, which is where the real problems begin. It is a first book, and as a first book in Spanish it is hard to beat; as your only book it runs out very quickly.