epub validator error

6 posts / 0 new
Last post

I have an epub book file created in InDesign. When I use the idpf epub validator it gives me the following warning. file: OEBPS/toc.ncx, message: meta@dtb:uid content ‘978-1-62397-028-4’ should conform to unique-identifier in content.opf:‘urn:uuid:978-1-62397-028-4’ Can this be fixed in the Indesign file? I do not know anything about writing code & any help would be appreciated.


The dtb:uid metadata field in the NCX file (the EPUB 2 table of contents file) has to exactly match the unique identifier in the package document. In this case, it looks like a bug in the InDesign export, as they should have included 'urn:uuid:' in the NCX file property (i.e., 'urn:uuid:978-1-62397-028-4' is not identical to '978-1-62397-028-4').

I'm not an InDesign user, but I'd be surprised if you can modify this field from within the program (only input the identifier for the ebook). The NCX would be generated on output, so you probably need to report it as a bug.

If you have a program that lets you edit files directly in a zip archive (which is all your epub file is), all you need to do is locate the ncx file (probably in an OEBPS directory), find the meta tag that looks like this:

<meta name="dtb:uid" content="978-1-62397-028-4"/>

and change it to:

<meta name="dtb:uid" content="urn:uuid:978-1-62397-028-4"/>

If it's an EPUB 2 file, Sigil might work for you to do this, but I wouldn't use that program on an EPUB 3 file (it's forced tidy of the markup is likely do terrible things to your html5).
You can also unzip the EPUB and modify the files instead of working in th zip. The only trick is that you typically can't just zip the contents back up again en masse, as there is a requirement that the mimetype file be the first in the zip container and uncompressed. The epubcheck program can recreate the file for you (on successful validation), but how comfortable you are with java and command lines will influence whether you want to go that route. See this thread for some more info.
Hope this helps.

I tried to submit an issue involving "Fragment identifier is not defined" but I got a message saying my "submission has triggered the spam filter and will not be accepted." How do I get around your spam filter?

Your spam filter is still blocking me. This is very frustrating.

I have 7 epub book files, with only the seventh having errors. The epub validator reports, "Fragment identifier is not defined" for two different locations: Line 264 in OEBPS/toc.ncx (a file that I cannot find, since I don’t know how to unzip the files on a Mac), and Line 91 in OEBPS/Text/contents.html, which I have found but cannot figure out how to fix.

Can you tell me how to fix these errors (and how to locate the first file)? The TOC link for line 91 seems to work just fine.


Assuming contents.html is your table of contents, it's probably the same error in both places.

Other than the fact that your table of contents is linking to a section that doesn't exist, or whose ID has been changed, it's impossible to say how to fix. You might need to remove the entry or change it somehow.

In terms of unpacking the epubs and fixing, have a look at this article:



Secondary menu