Data Migration Gotcha: Missing Authors

I have noticed that some migrated editions don’t have an author.

In the ogs era, some people, including myself used this approach: there was a Work type (somewhat similar to the Book type) which typically linked to the author. The author was then only associated with an edition, if mentioned in some other role, e.g. Copyright Holder.

As the Works haven’t been migrated, some editions are missing one of their most important bits of metadata. Also, they won’t show up under the Author credit list.

Recommendation…

  1. Always search by the book title (as well as author)
  2. Add authors to editions where they’re missing
  3. Review your previous submissions

Since those editions have to be edited, I really would like to be able to move the credits up and down. Because now, the author on said editions is at the bottom.

Ok, do you mean you want to drag and drop the credits? Like add a credit, choose Author, then drag it to the top of the list. I just want to make sure I understand this request.

@Burnett Drag and drop, that’s it. I just couldn’t pull that term out of my head yesterday.

It’d be useful to have the ability to re-order the main fields too, and have it preserve them in edits. Some fields keep their order, but some don’t, and some won’t let you order them with a raw edit.

Alternatively the data could be presented in a fixed, standard way that makes some sense. It’s better to have a logical progression of data presented on the Edition page to make it easier to read.

Agreed that having the ability to re-order credits would be great. Also images and URLs.

To be honest, that was quite disappointing.

I’ve started adding authors to my past contributions as instructed here, and noticed another thing that has gone wrong: on some anthologies and collections, the wrong Work has been migrated as the Book (usually one of the works included in them). I’m assuming these have to be fixed manually as well.

Can you give me some examples. I may be able to clean this up with a script.

When importing editions I made the assumption that an edition would always have a book and an author. So authors were skipped. I can restore those skipped authors on editions if a few of you think that’s a good idea.

:roll_eyes:

Have you ever tried at least once to type in info of a book you are reading right now?
Your assumptions make me wonder if you normally read…

Here you go:

Book: Lovecraft and I

Book: Rotkäppchen

Book: A Wizard of Earthsea

Because of this, collections containing the work can appear under the wrong Book page: A Wizard of Earthsea - Books - ctlgs

Plus, most of the magazines that include short fiction. It’s a lot. :sweat_smile:

But I want to note that it is completely possible that these did not have a “Book” to import. From what I remember, in the beginning there was some opposition of considering collections as Works, and we didn’t start adding them until later. I don’t think magazines ever had Works for the whole magazine, just the contents.

Yeah, I think it would be a good idea. It looks a bit weird if an edition doesn’t have an Author name anywhere.

Hopefully the ones that have already been added manually can be skipped.

Hey everyone, wanted to follow up on the issues raised in this thread. I’ve made three changes based on your feedback.

1. Missing author credits restored

I ran a cleanup script that restored author credits on editions where they were missing. If an edition was linked to a book with known authors, those authors have been added back to the edition’s credits. This covered about 36,000 editions. If you notice any editions still missing authors, let me know.

2. Anthology book links fixed

@auboisdormant — thanks for flagging the anthology issue. Editions like The Mammoth Book of Best New Horror were incorrectly linked to one of the contained works instead of the anthology itself. I’ve removed those incorrect book links on about 130 affected editions. You should now be able to link them to the correct book entry.

3. Credit reordering

@BadMoon — you can now reorder credits on the edit form using the up/down arrow buttons. No drag and drop yet, but you can move any credit up or down in the list. This works on all list fields (credits, dates, identifiers, etc).

Let me know if you run into any issues with these changes. thanks!

2 Likes

Thanks for all of these fixes!

I don’t think the order of fields is being preserved from a raw edit, though. I’m using that a lot as a form of copy-to-draft, and it’s usually easier to do edits right there and then. The order shown of fields on the Edition page of those Editions that I’ve submitted within the last hour (after you posted the above) aren’t showing in the order of the serialised raw data. Although, if I pick one and go to raw edit, the serialised data does seem to be in the same order I pasted & saved it. It’s just not displaying the fields in that order on the Edition page.

eg: https://ctlgs.org/c/books/edition/822602/scorpion-swamp/

Good catch. The credit roles on the edition page weren’t displaying in the same order as the data. This is fixed now — credits will show in the order you entered them. Your Scorpion Swamp edition should look correct now. Let me know if you see any other issues.

1 Like

Thank you for trying to fix them. But unfortunately, there are still plenty of them.

And that’s okay – might be easier (and more reliable) to fix them manually. Just going to take some time. :grimacing:

Ok, but if you see a pattern feel free to PM me and I’ll correct them. It is relatively easy to make bulk changes to the data.

1 Like

I just realized, editions not tied to a book still miss the author credit.

Even if the author’s credit page was imported, editions not linked to a Book have lost the author.
Creating a Book and link said editions to it doesn’t fix this (as I expected).

@Burnett So some author pages (some with actual profiles, not blank ones) weren’t imported since there were only editions tied to them but no Books. Is there a possibillity to import those in a batch or whatever it’s called? Since like it looks we have to manually edit the editions with no authors, it would be easier if the credit was in the database already.

BadMoon — yes, and it’s done. Those author credit pages are in the database now, and the editions have their authors back.

39,300 author credit pages that never made it across have been imported. These are the ones you identified: where an author was only ever attached to editions and never to a Book, their credit page didn’t come over at all.

70,933 editions have had their author credits restored.

The cause: the importer put authors on the Book and stripped them off the Edition, assuming every edition had a Book to hold them. It also linked each edition to only the first Work it pointed at. So an edition with no Work lost its author completely, and where that was the only place an author appeared, their credit page went too.

Two details:

  • Anything already fixed by hand was left alone — 308 editions.
  • Authors go at the top of the credits list now, not the bottom.

46,420 editions still show no author. I checked those against the old database and there’s nothing to restore — they had no author there either.

auboisdormant — on the anthologies linked to the wrong Book, I’m not going to run another partial batch after the last one left plenty behind. I found why it happens: where an edition listed several Works, the importer kept the first and dropped the rest, so a collection got linked to whichever story came first. There are 5,049 of them.

The dropped references are all still in the old database — 55,160 of them, covering around 33,000 Works that were never imported. That’s the same data as the Contents question xceque has been raising. So rather than patch the links, I want to bring those Works back and record the full list of what each edition contains. That fixes the wrong Book links, restores the missing Work pages, and gives Contents something real to start from.

If specific ones are blocking you meanwhile, PM me and I’ll correct them directly.

2 Likes

A follow-up on the credits from yesterday — two gaps I found afterwards are now closed.

Author photos. The credit pages I imported came in text-only. 6,638 photos have now been added to them, so pages like Alice Notley ( Alice Notley - books - ctlgs ) show their image again.

Other missing credits. It wasn’t only authors. 1,437 more credits of every role — publishers, illustrators, editors — were referenced by editions but had never been imported. Those are in now too.

Every credit referenced by a book or edition in the old database is now in ctlgs, apart from 7 that were broken over there as well: they sit in a credit slot but point at an edition record rather than a person, so there’s nothing to create.

Hopefully this covers all of the missing authors and original purpose of this thread. Let me know if you spot any problems.

1 Like