CSV Import Update: Custom Call Numbers

DiscussãoNew features

Aderi ao LibraryThing para poder publicar.

CSV Import Update: Custom Call Numbers

Este tópico está presentemente marcado como "adormecido"—a última mensagem tem mais de 90 dias. Pode acordar o tópico publicando uma resposta.

1kristilabrie
Editado: Dez 18, 2017, 3:48 pm

Thanks to LT dev ccatalfo, we've just added custom call numbers to .csv imports! Test it out with the updated "LibraryThingSample" .csv file on our Universal Import page: https://www.librarything.com/import.

Call numbers from the .csv import will get imported into the custom call number field, or "Other call number" field.

Though the import-export loop is still in need of being revamped/completed, this is a nice addition for those who need it!

2HighPrairieBookworm
Dez 21, 2017, 10:10 am

Can this be done retroactively? I have a boatload of books in my "Your Books" so that would be a nice addition to my list.

3lorax
Dez 21, 2017, 10:24 am

Oh, FFS.

The .csv import is deeply broken in profound ways, and you pick one random feature from one random request to implement, rather than just fixing the damn thing?

4kristilabrie
Dez 21, 2017, 12:36 pm

>2 HighPrairieBookworm: Unfortunately, no (you'd have to export/re-import, but I imagine that's far from ideal given the export-import loop isn't complete). Sorry for the trouble.

6r.orrison
Dez 21, 2017, 2:19 pm

I'd be pleased if they'd just fail the import if the column headers didn't match what they were expecting, instead of importing garbage.

7paradoxosalpha
Dez 21, 2017, 2:41 pm

>6 r.orrison:

That would be a great step. And I have to think it would be one of the easiest to implement.

8kristilabrie
Dez 21, 2017, 3:47 pm

>5 Petroglyph: I hear you loud and clear, it's definitely something that needs to be done and I know that devs know this.

9booktsunami
Dez 21, 2017, 11:55 pm

I am considering importing my library database (developed on Book Crawler) into Library thing. It presumably would be as a csv file but you say it is deeply broken. Tell me more. What is likely to go wrong for me. One thing I can't do apparently is import all my artwork associated with each listing. (And that is a problem for me). But what else is problematic. (I don't think i need "call numbers" anyway.

10lorax
Dez 22, 2017, 12:49 pm

There are a number of problems with the importer.

First, it will import only a very few fields, those that are available in the comically misnamed "Universal Import" example CSV. These fields must have exactly same names as in the sample file. Anything else will not import.

Second, If you have an ISBN in the file, it will happily overwrite title and author (not sure about date or publication) with information from whichever source you have chosen. If you haven't chosen otherwise, it will use Amazon which for older books is likely to be absolute garbage data.

Third, it will happily import garbage rather than warn you if something is badly formatted in your file, so people will get things like first-name only authors, or title and author swapped, for thousands of books if their file is messed up, and then give up in despair.

I've been begging for close to two years now for them to accept the full LT output as a viable import - basically, to allow import of every field that is user-editable - and while I get sympathetic noises, no developer has ever commented nor has anyone ever indicated it's actually even on the dev's radar so I suspect the non-technical staff is just patting me on the head to make me shut up.

11Petroglyph
Dez 22, 2017, 5:42 pm

Further to what >10 lorax: said:
There is no way of preparing in advance any book-level information that you want in in your LT catalogue, unless you're prepared to a) have the 'Universal Import' mangle everything that isn't isbn and then manually edit your copies one by one, or b) just manually enter everything. If you're particular about the metadata in your catalogue, a whole lot of manual editing is the only solution.

You can try to match the example .csv file for all fields except isbn, but then you'd have to manually add your isbn to all your items: as soon as an isbn is detected, the importer will scrape its source for the rest of the data; all the other info you prepared for that book will be ignored.

12PhaedraB
Dez 22, 2017, 7:01 pm

>11 Petroglyph: Also, when you do the import without ISBNs, the importer doesn't follow any instructions to add the import to particular collections.

13Petroglyph
Dez 22, 2017, 9:13 pm

>12 PhaedraB:
The more I find out about this import, the happier I am I no longer have any need for it. (Knock on wood against catastrophic data loss...)

14PhaedraB
Dez 23, 2017, 12:41 am

>13 Petroglyph: I transferred 1800 books from one account to another using export, import, and a lot of copy and paste. It was quite a project. Every time an improvement is made to the process, I get a little sad.

15bnielsen
Dez 23, 2017, 3:41 am

>13 Petroglyph: Also some of the fields that come from library sources are hopeless. I have 10+ books where the TSV export contains something that isn't unicode. It would be really nice to be able to edit those fields via an export/correct/import cycle. But as >10 lorax: says. It's not even "Pie in the sky".

16divinenanny
Dez 23, 2017, 7:32 am

>13 Petroglyph:
If I'd ever suffer catastrophic data loss here on LT I think I would just abandon LT altogether. Too much effort to get it back to what I need. I have my own DB in MySQL that is my master copy...