Dear Ferran,
Il giorno gio, 22/09/2011 alle 11.51 +0200, Ferran Jorba ha scritto:
> I've created an OAI set that contains records previously exported in
> another OAI set. I've noted that, at least in my current 0.99.1
> release, the OAI id appears only one, ex:
>
> 001 __ 200
> 024 8_ $9 artpubuab $a oai:ddd.uab.cat:200
> 024 8_ $9 driver
>
> I would say that the OAI id ($a) should appear twice, once per field, at
> least because if I delete the first OAI set (artpubuab) this id would
> not appear anymore in the record.
>
> I've digged a little bit into the code and the git history, and I have
> seen that it has been updated quite a bit, including a name change.
>
> What is the current practice? If the id should appear in each
> CFG_OAI_SET_FIELD (024, in our case), I'll try to patch our version as
> soon as possible, because my second OAI set must contain many more
> records and I cannot reindexing it more than once.
Indeed the Invenio implementation is somehow not perfect in the
treatment of OAI IDs, in particular with respect to multiple sets.
There is an on going effort to rewrite the OAI repository manager both
for speed and correctness.
In the future implementation there will be simply one field, with one
OAI ID, several current sets and several former sets (this is needed in
order to try to support the DELETED policy when a record is moved out of
a set).
IMHO, I would say the best practice to be future-compatible is to have
only one field with the only ID and repetitive $9 subfields for the
sets.
Cheers!
Sam
--
Samuele Kaplun
Invenio Developer ** <http://invenio-software.org/>
