Impossible to import an MARCXML bibliography

Eddy Louchez

Hello,

I will receive from the central service of my university a MarcXML file with all the books acquired during one year and useful for my bibliography.
When I will import into Zotero, I obtain the message "the file is not in a supported format".
I suppose it's a question of tags or maybe a COinS problem?
Here below the address of the file (an example):
http://bib.uclouvain.be/cgi-bin2/OAI2bis/VTLS/Vortex.pl?verb=ListRecords&metadataPrefix=marcxml&set=BTECNewBooks
I have tried to add the tag </collection> in editing the file in the notepad. Without result (I have an "import" but without any reference).
Anyone have an idea of the reason for the failure of import? Thank you very much by advance.

Eddy Louchez

noksagt

This has nothing to do with COinS. The main issue is that the returned response is MARC XML embedded in an OAI-PMH response. We can't process that yet.

Zotero's translators cannot currently strip the various response junk to get just the MARC data. An OAI-PMH translator might be possible in the future; it'd have to separate out the various record branches & send them to the relevant translator.

Importing this file "as-is" is not trivial, due in part to limitations in the Zotoer MARC XML translator. It is far too restrictive now & expects a single-lined marc:collection branch within the first five lines that is formatted in a very particular way. Even when forcing detectImport to trigger, I don't get a successful import, though. I haven't studied this in depth yet.

When I switch to mods, I can import. Sort of.
http://bib.uclouvain.be/cgi-bin2/OAI2bis/VTLS/Vortex.pl?verb=ListRecords&metadataPrefix=mods&set=BTECNewBooks

This yields another OAI-PMH response. So: you have to strip out that junk so you just have the content of the various modsCollection branches. But, again, Zotero expects all mods records (for each reference item) to be in the same modsCollection branch. So you further have to move some of those mods branches. This then imports just fine.

adamsmith

It's actually relatively simple to get the MarcXML to import. There are two issues:
1. The OAI-PMH header is too long, so the 5 line issue noksagt mentions kicks in. Not sure if we want to fix this on our end (checking more than 5 lines is costly). But simply deleting the header, empty lines, and response date and request at the top will fix this.
2. the namespace definition in the Zotero translator is too restrictive, so it doesn't detect
<marc:collection xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.loc.gov/MARC21/slim http://www.loc.gov/standards/marcxml/schema/MARC21slim.xsd" xmlns:marc="http://www.loc.gov/MARC21/slim">

only
<marc:collection xmlns:marc="http://www.loc.gov/MARC21/slim">


You can for now fix this with a simple search & replace, but we should fix this on our end.

noksagt

@adamsmith: FWIW, That's what I had done to get detectImport to trigger, but I still was not able to import anything. Did the multiple references actually import for you after you did this?

adamsmith

they did in scaffold, yes, but didn't try in Zotero.