search

4Science / DSpace

This repository contains the 4Science optimized DSpace & DSpace-CRIS distribution.
https://wiki.lyrasis.org/display/DSPACECRIS/
BSD 3-Clause "New" or "Revised" License
42 stars 61 forks source link

missing authority handling (COAR document types) in migration of dc.type #360

Open saschaszott opened 1 year ago

saschaszott commented 1 year ago

Bug description

In DSpace CRIS 7 document types are based on the COAR controlled vocabulary. This means that entries in metadata field dc.type are related to a COAR document type ID (taken from the IDs in file publication-coar-types.xml). The COAR ID is managed in column authority of the metadatavalue table.

The migration procedure of metadata field dc.type does not consider the handling of authorities. The values of document types (dc.type) are copied from the old DS CRIS 5 without any mapping to metadatavalue.text_value. The additional lookup of the COAR document type ID is missing.

After the migration the entries in dc.type are not connected to COAR types (column metadatavalue.authority contains NULL values).

In the end, this leads to a situation where duplicated document types arise, e.g. in the doc type browsing view (see screenshot). The authority value of dc.type in new submissions is set correctly, the migrated publication items have the NULL value in metadatavalue.authority.

image

To fix this bug an additional lookup step is required.

Expected behavior

Column metadatavalue.authority contains COAR IDs (expected syntax publication-coar-types:ID).