Russo and Ferrari wrote the reference for the calculation language behind Power BI, Excel and Microsoft's analysis services: twenty chapters running from the basics to query plans. What I read is not that book but a commercial summary that walks it chapter by chapter in tables, with a section of quotes and a quiz at the end.
What I took from it
A map of the terrain, and confirmation of where the difficulty sits: evaluation contexts, the transition between them, and the function that rewrites them. The summary names those topics and orders them, which is enough to know what has to be studied and how much time to ask for.
And a lesson about the format that outlives this case. A summary keeps the argument and throws away the examples. In an essay that can work, because the argument is the book. In a technical book like this one the argument is trivial — the filter context decides the result — and all the value lives in the dozens of worked examples where you see why your measure returned a number you did not expect. Without those examples you are left with vocabulary, not knowledge.
The closing quiz makes it obvious: true-or-false statements, several written to be false, and anyone who has not read the book has no way to tell them apart. A summary that can teach you something wrong without your noticing is worse than reading nothing.
What I left out
The book, which is the point, and that is why this entry has no opinion on it. What I will defend is the criterion: for technical material, summaries are only good as an index of what you will have to study anyway. This one is pending, in full, before I say a word about optimising a measure.