You can specify what metadata to download. For new books with very bad metadata I usually set ISBN and download everything else. And then carefully normalize what is downloaded to what already is in the library. For books already in the library I usually only download comments and cover.
The series information that is available for download usually is very unreliable. The reason is that there usually are several different possible series names as well as several different series numbering possible. Order of publishing or internal chronology or suggested reading order by the author or by some fan. Also the spelling of the authors may vary. Minor things like different punctuation and initials for middle names may cause calibre to move a book to a completely new author. Even the title can vary. Junk like "A Novel" or "A NN book" or other things may be added.
It's up to you to keep the metadata normalized.
Download and normalize author, title and series information when the book is first added to the library. And at that time make sure the spelling of the authors is the same as the spelling used in all other books by the same author. And that you use the same series name and numbering. And so on... Don't download that metadata again, because it would most likely ruin your normalization.
It might help if you only download metadata from only one source. But even then things like series information may be bonkers. Junk in, junk out.
|