When downloading metadata, does it merge all sources or use only one source?

[Lily7]

When downloading the metadata from various sources, does calibre put in all the metadata gathered from all of those sources, or does it only use one source and if so how do I know which source it's using?

For example, say Amazon has 5 tags for a book and Goodreads has 10 tags, does it only put Amazon's 5 or Goodread's 10 or does it put up the duplicates and any extras, merging the two?

Similarly, how does it deal with ratings? Does it average out the ratings from all the sources or only select ratings from one source?

Etc. You get the idea. Thanks!

[kovidgoyal]

It uses a complex algorithm that selects the "best" values for the same book from different sources, sometimes merging, sometimes overwriting. The only comprehensive answer is to read the source code to see exactly what it does in metadata/sources/identify.py

[DoctorOhh]

Originally Posted by Lily7

For example, say Amazon has 5 tags for a book and Goodreads has 10 tags, does it only put Amazon's 5 or Goodread's 10 or does it put up the duplicates and any extras, merging the two?

I only have calibre grab tags from Goodreads. Of the metadata sources I have configured, Calibre can get tags from Google, FictionDB, Overdrive or Goodreads. In the metadata download page I have configured each Google, FictionDB, Overdrive and unchecked Tags so the only source left for me is Goodreads.

Originally Posted by Lily7

Similarly, how does it deal with ratings? Does it average out the ratings from all the sources or only select ratings from one source?

If you prefer a particle source for downloading ratings simply configure all of the sources you use and uncheck ratings for the sources you don't want to use for ratings.