Hi Tibor,
 
> On Fri, 01 Apr 2011, Ferran Jorba wrote:
>>   duplicated = 27
>>     oai:www.recercat.net:2072/88000 = [63044, 67648]
>
> Looking at these two records on your site:
>
>   <http://ddd.uab.cat/record/63044/export/hm>
>   <http://ddd.uab.cat/record/67648/export/hm>
>
> I can see that they have fully identical 035 values.  BibUpload should
> not allow you to create a duplicate record in this case; rather, it
> should end with an error if you try to submit the latter record as new.
> (This behaviour is governed by CFG_BIBUPLOAD_EXTERNAL_OAIID_TAG.)  Have
> you looked at why it created a dupe for you?

Yes, I have ;-( We are moving from 970 to the more standard 035 for
external identifiers (I backported
http://invenio-software.org/repo/invenio/commit/?id=2fb7275849e83f5afbb7915000e208a3e053889a
to 0.99.1), but old converters pop up from time to time still using 970.
And I noticed this site was one of them, and I corrected it during my
debugging sessions and correcting them afterwards.  So in this case it
is just my fault...

> Also, if you are using OAI harvester with Invenio v0.99.1, please
> consider upgrading to v0.99.3.  (The upgrade from v0.99.1 to v0.99.3
> should be straightforward.)  There have been several bug fixes in the
> OAI harvester such as:
>
> <http://invenio-software.org/repo/invenio/commit/?id=b4e6b792e942a5c38149fb218c2b50ca50c898ac>

Ooops!  Thanks for remembering it.  I'll *really* have to do it...

Best regards,

Ferran

Reply via email to