Where are metadata text_lang defaults determined?

[Fitchett, Deborah]

When we (v5.8 XMLUI Mirage2) export a CSV from Batch Metadata Editing, we can get multiple columns for each metadata field due to different “text_lang” codes being assigned = eg dc.title, dc.title[], dc.title[en], dc.title[en_US], dc.title[*], etc.  This is partially described at https://wiki.lyrasis.org/display/DSPACE/TechnicalFAQ#TechnicalFAQ-MetadatavaluesinCSVexportseemtohaveduplicatecolumns and I’m comfortable with using both the CSV and SQL fixes for existing data. But I’d like a more permanent solution for data created from this point forward which, as the page notes, depends on the import methods.

 

We have data from the following sources:

1. Records already in DSpace when we use the Edit Metadata functionality – this mostly does what we tell it transparently, but experimentation shows that on saving, any null language codes get saved as '' instead. I suspect this is hard-coded and we’ll just need to live with it.
2. Records manually submitted direct to the DSpace workflow – these seem to automatically add our instance’s default language en to some fields; * to the Creative Commons license fields; and null to other fields. Is there somewhere we can edit any/all of these defaults?
3. Records submitted from Symplectic Elements via SWORD v2 – these come across as null, except a dc.provenance field is created by DSpace with en. We’re checking with the vendor whether we can specify language codes on their side, but I thought I’d better ask on the DSpace end are there any configurable defaults in SWORD?

 

Our dspace.cfg file has

# Default language for metadata values

default.language = en

 

but this doesn’t explain why only some metadata fields are using that value while others are using null or *. I can’t find anything in any of the other config files or through the web interface. Is there anything configurable, or is the ‘fix’ to just periodically run either the CSV or SQL fix?

 

Deborah

––––––––––––––––––––––––––––––––––

Deborah Fitchett (she/her) MLIS, RLIANZA

Associate University Librarian, Digital Scholarship

 

––––––––––––––––––––––––––––––––––

Learning, Teaching and Library – Te Whare Pūrākau

PO Box 85064, Lincoln University

Lincoln 7647, Christchurch, New Zealand

+64 3 423 0358

deborah....@lincoln.ac.nz

ltl.lincoln.ac.nz

 

––––––––––––––––––––––––––––––––––

Lincoln University

Te Whare Wānaka o Aoraki

––––––––––––––––––––––––––––––––––

 

---

"The contents of this e-mail (including any attachments) may be confidential and/or subject to copyright. Any unauthorised use, distribution, or copying of the contents is expressly prohibited. If you have received this e-mail in error, please advise the sender by return e-mail or telephone and then delete this e-mail together with all attachments from your system."

[al...@vt.edu]

We also have "en" as our default language code.  "*" and [] are added as you describe. And we get deposits from Elements as you describe. Our solution has been a nightly script to convert everything to [en], as reported in https://github.com/VTUL/vtechworks/issues/737.

[Fitchett, Deborah]

Thanks, it sounds like there’s no control over the language codes from the Elements end, so a regular script seems like the easiest way to go.

 

Deborah

 

From: dspac...@googlegroups.com <dspac...@googlegroups.com> On Behalf Of al...@vt.edu
Sent: Friday, 22 April 2022 4:39 am
To: DSpace Technical Support <dspac...@googlegroups.com>
Subject: [dspace-tech] Re: Where are metadata text_lang defaults determined?

 

Caution: This email originated from outside our organisation. Do not click links or open attachments unless you recognize the sender and know the content is safe.

 

--
All messages to this mailing list should adhere to the Code of Conduct: https://www.lyrasis.org/about/Pages/Code-of-Conduct.aspx
---
You received this message because you are subscribed to the Google Groups "DSpace Technical Support" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dspace-tech...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/dspace-tech/c252f0c8-b036-4e08-9fca-a8f0416e22e3n%40googlegroups.com.

[Anne Lawrence]

Yes, all of us who deal with metadata here breathed a huge sigh of relief when the script was implemented. CSVs with 20 columns are a lot to handle. But they sure beat the same ones with 50+ columns!

Anne

--

Anne Lawrence

Repository Application Administrator

Virginia Tech | University Libraries (0434)

Newman Library, Room 420

560 Drillfield Drive

Blacksburg, VA  24061

(540) 231-9320 | annela...@vt.edu

[Fitchett, Deborah]

One quirk to be aware of for any places that do this:

 

The metadatavalue table holds metadata for *everything*, including bitstreams, communities, epersons, etc, not just bibliographic metadata from items. It turns out (at least in our 5.8 instance!) that if you set a non-NULL text_lang value for a bitstream’s metadata value, then when you try to manually edit the bitstream name or description it appears to refuse to save.

 

More precisely, I think it creates a duplicate row in metadatavalue with the edited value, but the old value is the one that continues to display in the user interface. I didn’t delve into this too deeply but/because setting all of the text_lang codes back to NULL where resource_type_id='0' solved what had otherwise been a very mysterious problem!

 

So if/when we set up a script to make changes regularly, we’ll be limiting it to where resource_type_id='2'  (ie items only).

 

(Resource types: https://wiki.lyrasis.org/display/DSDOC5x/Business+Logic+Layer#BusinessLogicLayer-Constants )

 

Deborah