Hello,
when importing from external OAI sources, I'm finding records empty
fields, for example:
<datafield tag="500" ind1="" ind2="">
<subfield code="a" />
</datafield>
On those cases, both xmlmarclint and bibupload complain and the whole
procedure fails, including bibsched getting stopped, etc. If this
happens only occasionaly, afterwards I take care of those errors
manually and fix them. However, my understanding is that this field
should quietly disappear when imported to my site.
However, now I'm considering importing from a number of OJS systems
(http://pkp.sfu.ca/?q=ojs) that systematically export Dublin Core
records with empty fields, like the example above.
I'm seeking some advice about how, where or when to deal with them.
Should it be done just during the Dublin Core to Marcxml conversion
(say, etc/bibconvert/config/ojs2marcxml.xsl) or in the MarcXML parser
(lib/python/invenio/bibrecord.py), in the general function
(create_record) or for each of the low lever parsers (create_record_RXP,
create_record_minidom create_record_4suite).
I'd say that the later solution is better, because it would provide a
more general solution, but I feel somewhat intimidating.
Is there any progress done in this front that I can backport or word of
advice?
Thanks,
Ferran
