Having started reading QFT as an undergrad from textbooks like Mandl and Shaw, I was reluctant to use this one, even while it was the recommended textbook of a graduate course in field theory. The main reason for this was that Peskin and Schroeder (P&S) makes practically no effort to make contact with the rest of the (vast) literature on the subject. If you have read some other QFT book it is very-very difficult to go through P&S and vice-versa. I remember trying to use in some occasions this book for some calculation and ending up completely confused, because the notation and normalization conventions where different from everybody else. So after these first sad encounters I quickly dismissed it and decided to use other books for QFT instead.
Unfortunately, P&S seemed to remain the standard reference and everybody else seemed to have read it, so from some point on, I decided to give it another chance, so I wouldn't feel I was intellectually isolated. Thus, I bought the book and spent a couple of months reading through most of the text. This time I decided to not just read the parts I considered new, but start from the very beginning and keep going, doing every in-between calculation. Surprisingly, this time I could understand what was going on and managed to advance very fast through the chapters.
I realized though that my initial impression remained true. The book is very idiosyncratic in its presentation method and many topics are treated here in a way you won't find anywhere else. This can be actually very useful, if you have already some familiarity with the material and you want to gain some further insight.
The chapters of P&S have an obvious flaw though, which is why I couldn't follow the text on my first attempt: They are not at all self-contained. The book will present some small, one paragraph argument, which at the particular point seems rather tangential to what you are reading, then 400 pages latter, in a different chapter and subject, there comes a reference to that argument which now appears to be of outmost significance. So, you have to go back and see what is it that you missed. Apparently, unless you are reading the book without stop and start to finish, there is no way to avoid these frustrating self-references (and even if you are reading full-time, it takes about two weeks to advance 400 pages and by that time, you have most certainly forgotten half of the things you 've read). Many chapters suffer from the same problem and this renders the book almost useless as a reference, Every time you have to look up something which is a little more advanced than the Dirac equation, you end up encountering some reference to a previous passage, which then references another and so one, until you have to read again half of the book to find what you where looking for.
There are also parts where an argument on a subject (like the Ward-Takahashi identity) can extend through many chapters and many pages. It is not uncommon in P&S to find discussions which continue for more than 10 pages. By the time you reach the end, you have almost forgotten what you where trying to prove in the first place. And this is another problem of the book. It has a tendency to present subjects which are in fact difficult and obscure as long discussions, without giving a hint in the beginning about what the result will be and expecting from the reader to make up his own mind about what actually has happened over the past 10 pages. Even when the exposition is interesting and engaging, it still may leave the reader perplexed in the end. The book also makes no distinction between which parts are "considered" easy and those that are supposed to be more difficult. This is very frustrating for the reader, since he may end up struggling too much over an easy part for no reason, then the next moment not paying the attention needed to truly follow a more profound section. It is always easier to learn once you are told what to expect.