Kobo confused by duplicate book

[franklekens]

Okay, here's something that seems to go wrong more than once when I sideload books with Calibre. I remember it went wrong once or twice before on my Forma, more or less exactly the way I describe below, and I've just encountered it again. I'll try and describe it as clearly as possible.

I had a free copy of Mark Twain's Innocents Abroad that I downloaded from Munseys.com (a now defunct website with copyright free ebooks). This was in my Calibre library on my desktop (Windows 10), and it was on my Forma.

I don't know whether this also matters, but I'll mention it just in case: I had opened this e-book on the device, and it was the last opened book, it was the first on the home screen.

Now I happened to download the Gutenberg version of the same book today, it seems better formatted and it has illustrations. It's a huge file, however (8 Mb). I added this to Calibre, I got the warning about duplicate books and I clicked okay. I now had two separate copies of this book in Calibre.

I connected the device to the computer. In Calibre it showed that the Munseys copy of the book was on the device (check mark), and the Gutenberg one was not as yet (no check mark). I uploaded the latter to the device, without removing the former. (I know: why keep two copies? But never mind. My point is: I should be *able* to, surely? If I wanted?)

Okay, so I disconnect the device. The Forma shows me the home screen. I already sense something is wrong: as the last opened book it does show Innocents Abroad, but the image I see is that book cover that comes with the Gutenberg version (https://www.gutenberg.org/ebooks/3176). And in the book list it showed the size as 8Mb, not the appr. 500 Kb that is the Munsey version's size.

So Calibre may have added the second book to Kobo, but either it hasn't made it clear to the Forma that this is a second copy of the book that just happens to have the same title or (more likely) the Forma's firmware or database structure just can't handle it and gets confused. Because when I tap this book and try to open it, it says "oops, the book can't be found".

It's quickly solved, that's not my problem now. That is to say, I delete the book through the device interface, rebooted it just to be sure, reconnected it to the computer, Calibre now showing no copies of this book on the device, uploaded the Gutenberg version -- and now it's on there and there are no problems.

But surely this is a glitch in the way the Kobo devices (sometimes! -- it doesn't seem to go wrong all of the time) handle duplicate titles when sideloaded via Calibre? Maybe something can be done to remedy this?

(What I didn't do, is reconnect before deleting the book and see what actual files were on the device. Nor did I check whether I could reproduce it by going through the same steps, first uploading the Munseys copy &c. If someone really wants to I could try and see if it's consistent. Though I hope it's not necessary, I'd just like to get reading. Just thougth I'd describe what went wrong in case anybody else encounters it or wants to investige this to see if it can be remedied.)

[PeterT]

When calibre adds a book (by default) I believe it uses a location based on the Author and book title, so if it is a duplicate calibre would just overwrite the existing book. (I'm willing for this answer to be shown as false! I'm not at my PC right now/.

Sent from my Pixel 4a using Tapatalk

[milady133]

I also suspect that what has happened is what @PeterT said. The first book was substituted with the second one. Calibre plugins update Kobo's internal database with the metadata of the new book, but the internal structure of the book is also on the database, and Calibre doesn't update that, so the new book has a different internal structure and Kobo can't match the internal files of the old book with the ones in the new one.
In these cases, when I have a duplicated book in my Calibre library, what I do is modify the title of the book so I can see quickly the reason to have them duplicated. For example, in my case I have two versions of the book Arabella by Georgette Heyer, one in Spanish and one in English, so my books have the title "Arabella - Español" and "Arabella - English".

[DNSB]

I am wondering if the OP has the "Do not treat replacements as new books" option checked in the KoboTouch or KoboTouchExtended driver configuration. It's in the second tab "Collections, covers and uploads".

[davidfor]

Kobo ereaders have no problem with having multiple copies of books with the same metadata. They differentiate the books by the file path. All you need to do is change the filename or directory somehow and the device will have no trouble differentiating the book. Whether you can remember which is which, is a different matter.

So...

What is almost definitely happening is that when the second edition was sent to the device it replaced the original copy. That will be because the two copies have the same title and author. When a book is sent to the device, the save template that is defined in the driver is used to set the file name. The default for the save template is something like:

{author_sort}/{title} - {authors}

That means if calibre has to copies of "Some book" by "Some Author", when either is sent to the device, the book will be named "Author, Some/Some book - Some Author". The second copy will silently replace the original copy. This is working as designed.

The solution is to either change the metadata that is used in the save template so that they are different, maybe "Some book - munsys.com" and "Some book - PG". That would also make it easy to work out which book it is you are looking at on the device. Alternatively, you change the save template to use something that will generate a different filename. The simplest is to use the book id that calibre has. You could use:

{author_sort}/{id}{title} - {authors}

That is guaranteed to be different for books in the same library.

Or even just:

{id}

Again that is guaranteed to be different for books in the same library, but, if you used multiple libraries, it could cause problems. Of course, if you use this, all books will be in the root directory of the device. That probably won't cause problems unless you have a lot of books. But, I would not like it.


There is an issue with having books with the same title and author on the device. But, that is with calibre. Calibre matches books on the device to the library using the title and author. That means if you have to copies of the book in the library, calibre does not have a way to work out which to match to the book on the device. And if you have both copies on the device, they will probably be matched to the same copy in the library. You might be able to change this with manual matching, but, I have never tried.


If you are changing the copy of the book on the device, there is one more gotcha. When the device imports a book, it imports the metadata and ToC from the book. If you updated the book on the device, it's default behaviour is to see that it is different (only using file size), remove the book completely from the database and then import it as a new book. That means you lose the reading status, annotations and collections. The driver in calibre can bypass this so the book is not treated as new. That retains the book status and works fine if the changes do not change the internal file or ToC structure. Fixing spelling, styles and layout work fine. Changing the ToC doesn't. But, my Kobo Utilities plugin has a function to do if you are updating kepubs or epubs. It doesn't try to fix the annotations, but, if you are changing the ToC titles, or adding chapters (useful for fanfiction and similar), it will work.

[franklekens]

Thanks for the explanation. The save template was indeed set to the default, I'll have a think whether I want to change that.