Hello David,
Quote:
Originally Posted by dpierron
Well, up till now the author_sort field was set on the BOOK (and not AUTHOR) table, which could lead to a single author being spelled differently (and thus referenced more than once). This has changed recently, IIRC, I'll have to check.
|
author_sort is still in the book, as it always was.
Quote:
|
Also, this field usually contained garbage for most users, who leaved it as is without even checking what Calibre did with it (and if you imported an ePub file, it would take the content of the ePub metadata for hard information... dunno if this changed).
|
Nope. Author_sort can (and often is) filled with bizarre stuff.
Quote:
|
So yes, it's bad and no, I will not change it lightly; but yes, I'll try and remember to have a look into the new Calibre features regarding this issue.
|
The relevant change to calibre is that each author now has an individual sort string, stored per author, not per book. When the user asks that author_sort for a book be filled in (pushes the button in edit metadata), calibre goes to the author table to get the individual sort strings.
As C2O uses the db directly, you can ignore the book's author_sort field when generating author lists. Instead, use the sort value in the author table. That will eliminate all the parsing ambiguity and the multiple-value problem.