Metadata changing information in Calibre

[Dragonhomer]

I wrote about this before, and the answers were mainly "why would you want to do this" or "books have the metadata in them already" or "why are you downloading so many at a time"

First, I want the metadata so I can have an idea about the book - just like I used to do when I went to the bookstore to buy a book. I would read the back cover to get an idea about the book.

Second, a high percentage of the books I download DO NOT have the metadata attached.

Third, I download a lot of metadata at a time so I can do other things while it is downloading. I have limited time to spend on the computer because I am a caregiver for 2 ladies and don't have a lot of free time during the day. In the evenings I perfer to be reading rather than downloading metadata.

Today I had a day off and decided to spend time trying to get more details to pass on to you to get more help (hopefully). This is what I found out:

After making sure the 20 books I was going to get the metadata on didn't have metadata already, I made sure the title, author and cover were correct. Then I ran the search for metadata.

Out of the 20 books:
7 books came back with all the title and author information changed and the metadata for those books in the comments section.

A year ago, I didn't have any of these issues - I used to download the information and other than an error once in a while. Now it seems like I am having to doublecheck the information in every download.

I do not understand why the companies providing metadata for people to download put blocks in our way or limit us to the amount of information we can get.

I am at the point that I am going to write to Amazon (where I get about 75% of my books) and complain to them about it. Maybe losing one customer won't bother them, but if I post about this on the groups I am a part of, maybe other people will write & complain as well.

I am wondering if something in the program changed or did the companies providing metadata just decide to start messing with the information people are requesting.

Any constructive suggestions?

[kovidgoyal]

Nothing has changed in metadata download in calibre in forever. Quality of downloaded metadata is entirely a function of the sources you have enabled for metadata in calibre. There are a large number of third party plugins for metadata download some of which might work better for you than the builtin amazon/google ones depending on the types of books you have.

And just FTR there are no companies that provide metadata for users to use. Indeed many of these companies go out of their way to block people from using their databases for metadata, amazon being the prime example. The calibre amazon metadata plugin actually doesnt query amazon at all, it goes via google search of amazon and then uses the google webcache of amazon to actually get the data, and has done so for many years.

[sgmoore]

I have the same issue and unfortunately no real solution other than always use the 'Review Downloaded meta'

The problem is compounded by the default titles I get when importing amazon books. For example I have one book with the title

The Poison Path: A Gripping Detective Crime Mystery (The DI Hogarth Poison Path Series Book 1)

When I look for metadata from google what it actually sends is a request

https://books.google.com/books/feeds...ewability=none

So it is looking for 15 things in the title and doesn't find a match but does find several other books in the series than match 12 of those things and thinks that is good enough.

Hence it suggests a completely different book.

Usually I get better results by changing the title to be just "The Poison Path" , but in this particular case I get a book by a completely different author.

I've tried excluding google from the metadata sources , but that means I often don't get any results.

If I'm correct there isn't any way of prioritizing the meta data results as there does when download covers.

[kovidgoyal]

No there is no prioritization for non-cover metadata sources. Indeed, metadata from multipe sources is combined based on isbn/title/author matches. That way the actual end results are much more likely to have all fields filled rather than only some fields depending on what metadata a particular source returns. As such prioritising them doesnt make sense.