I read this book end to end, learning something on every page. It is full of great tricks from recursive queries, how to use the Alerts log as an External Table so you can query up-time, how to find un-indexed FKeys and why they matter, to basics like table types, column types, index types (when would you use a reverse index?)... Did you know that Kyte never uses CHAR? What field type should you store manually encrypted values in and why?



There are so many useful bits of information in this book. List of when not to use a cluster (need to truncate?, need to full scan?, heavy updates?, need to partition?). Types of partitioning and when to use them. Types of encryption and when to use them (sometimes encrypted data is encrypted in memory as well, and sometimes it isn't. Is encrypted data always stored as encrypted when written to disk (Redo, Undo, Temp, Rollback)?). What features can Exadata apply at the IO level (column filtering, row filtering, decryption, regex, etc). How's your parallel DML?



Kyte is not just great at Oracle. He's a great teacher. Despite the dry subject, Kyte makes it possible to take in what he has to share.



Recommendation: begin this book by installing Oracle and enabling yourself to follow the examples. Kyte walks you through how to set it up before page 1.