Why tags download only after second click on "Download metadata"?

[fufu42]

Hi,

whenever I download metadata for some new books, the tags are only downloaded after I proceed a second time through the "Download metadata" process of selecting edition and selecting cover. Is there any intention in this or does a setting exist?

Thanks
fufu

[kiwidude]

Only some metadata sources (such as Google I think in the default calibre plugins) support tags. You can download others that do (and provide "genre" tags rather than "random stuff" like Google keeps) from the Plugins subforum, such as the "Goodreads" plugin.

When a metadata download search takes place, the results are grouped into different ISBN pools based on the first match returned from each website. If those websites return different editions (or completely different books if their search engine is particularly crap) then calibre displays each group of results as a row in the second page of that Metadata Download dialog. If you are using Bulk download, then the first row will automatically be chosen - which may or may not be the right book, and will certainly only represent a subset of those metadata sources (so Google might be in the second row of results, and hence not chosen so the tags don't appear).

The second time you do a download, your book will have an ISBN allocated to it from the first download (amongst other ids). This gives the metadata search engines a better "hint" as to which book you are after, and increases the likelihood that more of the metadata download sources end up with the same edition and hence same first row of results. So in your case it may mean the Google result now ends up as the default selection, and hence the tags appear.

So those of us who are particularly picky about the quality of metadata frequently require "several passes" to get all the results from all the desired website sources. Sometimes you can never get them to appear in the same one group, due to the site not stocking that book or edition that was chosen from another site.

If you want to improve the chances of getting common data in a single pass, make sure your book has an ISBN. If you don't have one from the metadata, you can use the Extract ISBN plugin. Of course that plugin too could be fooled by the contents inside the book (e.g. an advertisment for another book by that author/publisher) or the book may not have an ISBN.

So it is a case of trial and error, watching the results etc if you want to make sure you don't end up with nonsense metadata...

[fufu42]

Thank you for this quick and insightful response.