Calibre is wrong about which versions of a book are on my device

[Siavahda]

I recently imported one of my back-up libraries into Calibre, which might have caused this problem? Because now there are duplicates of many books in my Calibre library (usually just two versions total).

And Calibre is incorrect about which ones are on my Kobo. It claims the AZW3 versions are on my Kobo, which is obviously not the case. (I did triple-check, but no, all the files on the Kobo are EPUB, just like they should be.)(Specifically EPUB, not KEPUB, in case that's relevant.)

The versions of the books-on-Kobo are still in Calibre, they haven't been removed or the metadata edited or anything. I don't know why Calibre's not recognising that these are the books on the device.

Is there any way to make Calibre (or the Kobo) do some kind of sync to correct this? (I did try the 'update cache metadata on device', but that didn't help.) Currently the Kobo isn't displaying things like the correct series metadata because that information isn't attached to the versions of the books it thinks (or Calibre thinks?) are on there.

I don't know the right terminology for any of this so I hope this makes sense to those better-versed than me!

[Siavahda]

Deleting metadata.calibre off my Kobo caused Calibre to re-scan, but it still thinks the wrong versions of the books are on there.

[Siavahda]

Deleting the AZW3-only folders from the Calibre library, deleting metadata.calibre off the Kobo, and having Calibre re-scan again makes it recognise the actual, remaining versions of the books as the correct ones.

But there's 10k+ books, I can't do them all one by one. Hope there's a quicker fix!

[jhowell]

calibre matches books on a device with your library based on title and author, not format.

[Siavahda]

Originally Posted by jhowell

calibre matches books on a device with your library based on title and author, not format.

I didn't know that. Does that mean there's no fix? But wouldn't that mean that this problem would arise ANY time you had duplicate books? This hasn't happened to me before and I've had duplicates in my library for years.

Is there a way to get Calibre to match the correct books with the device?

[theducks]

Find Duplicates Plugin can locate when there is more than 1 entry in the Library.
You can have ONE of each format IN an line entry., so you just merge the 2 records 1)select the target. 2)hold the control key and select the 'to merge' record(s), 3)tap 'M'
Note merge does not replace existing metadata values, only adds that when the value is blank.

[Siavahda]

Originally Posted by theducks

Find Duplicates Plugin can locate when there is more than 1 entry in the Library.
You can have ONE of each format IN an line entry., so you just merge the 2 records 1)select the target. 2)hold the control key and select the 'to merge' record(s), 3)tap 'M'
Note merge does not replace existing metadata values, only adds that when the value is blank.

Sorry if I'm being stupid, but where is the 'to merge' option? Nothing appears when I hold control after selecting the two books/entries.

Also, this sounds like I'll have to manually go through and do this for every entry. Is that right?

[theducks]

control to select the second and/or more books. the release.
M is the keyboard shortcut
'Merge' is not a context menu option to add.

[JSWolf]

You can auto merge on adding books.

This is from Gemni AI and it's correct. I double checked.

To automatically merge the same books of different formats together upon adding, configure Calibre's auto-merge preferences:

1. Click Preferences > Adding books (under Import/export).
2. Go to the Adding actions tab.
3. Check the box for Auto-merge added books if they already exist.
4. Choose your preferred merge action from the drop-down menu (e.g., overwrite or keep unique formats), then Apply.

[Siavahda]

Originally Posted by JSWolf

You can auto merge on adding books.

This is from Gemni AI and it's correct. I double checked.

To automatically merge the same books of different formats together upon adding, configure Calibre's auto-merge preferences:

1. Click Preferences > Adding books (under Import/export).
2. Go to the Adding actions tab.
3. Check the box for Auto-merge added books if they already exist.
4. Choose your preferred merge action from the drop-down menu (e.g., overwrite or keep unique formats), then Apply.

That looks like it would work, I'm just hesitant to do it. Mostly because I'm not sure if I understand the best option for my library.

If I could set it by DATES, maybe it would work - any duplicate from after the 1st of this month can be overwritten. Trying to do it by format... There's a lot of duplicates that have the same formats, but different ages. And I think it's case-by-case whether the older or newer version is one I'd want to keep; sometimes I've gone back and fixed typos in a book I've read, and that's not necessarily the most recently modified version, because what's come in from the backup a few days ago (which Calibre reads as newer, obviously) has the older text in it, with the typos unfixed. Does that make sense?

If I have two duplicates, and they each have, say EPUB and MOBI in them, how does it decide which gets overwritten? If the files in one duplicate are newer, or have been modified more recently?

[JSWolf]

If you have duplicates of the same format but different dates, add them to your library and all the library to import the duplicates. Then use the Find Duplicates plugin to find the duplicates based on the same/author and then you can choose which to keep and which to delete.

[theducks]

Originally Posted by JSWolf

If you have duplicates of the same format but different dates, add them to your library and all the library to import the duplicates. Then use the Find Duplicates plugin to find the duplicates based on the same/author and then you can choose which to keep and which to delete.

What 'date' do you use? (AFAIK inside Calibre, The file date (the only built in way to see that is to hover over the format in the MDE) or the Pub date?

[JSWolf]

Originally Posted by theducks

What 'date' do you use? (AFAIK inside Calibre, The file date (the only built in way to see that is to hover over the format in the MDE) or the Pub date?

I usually go by the date added in calibre. I have a library that's my source library. If I want to read something, I'll copy it to my working library and edit the eBook from there.

However, there are other solutions if they two copies are not binary duplicates.

I use the TextDiff plugin. That lets me know if there are differences in the text. I can then decide which I want to delete. I know the calibre editor has an eBook compare feature. But a lot of the time it doesn't work because the filename have been changed. SO it won't be able to compare if that happens. But TextDiff will still be able to compare.

Another thing that works is some publishers have a version number in the copyright page. So you can have a look at one of the copies and see if it has a version number and if so, take the one with the later version.

[maddz]

How does that work with books of the same title but different authors? I'm thinking of the various short story collections from Standardebooks which are all called 'Short Stories' or translated works with different translators; I want to keep those separate. I know I'm always running foul of the add books dialogue which says 'you already have this in your library'.

[DNSB]

Originally Posted by maddz

How does that work with books of the same title but different authors? I'm thinking of the various short story collections from Standardebooks which are all called 'Short Stories' or translated works with different translators; I want to keep those separate. I know I'm always running foul of the add books dialogue which says 'you already have this in your library'.

I modify the titles most often I use the publication date (Short Stories, March 2026). For translated works, I add the translator with ,trans as a second author.