Re: [Bug 516] biblioitems.itemtype does not load
The "only solution" is not a solution at all. The only place the itemtype exists in a MARC record is in the holdings tag (be that 952, 995, whatever), so Koha _must_ be able to move that info into the biblioitems table. The user should not have to invent a workaround!! Just a general note -- DATA IMPORT PROBLEMS ARE CRITICAL!! The libraries you want to see using Koha are already going to have their data in a commercial ILS, and Koha must be able to smoothly incorporate that data into its tables. I appreciate the beauty of the Koha database structure, but the import scripts cannot be limited by that structure. Since the itemtype info is located in the MARC holdings tag, then Koha has to retrieve that info from the MARC tag and put it in the appropriate table. If items attached to a MARC record have different itemtypes, that is a MARC cataloguing error and should generate an error message when an import is attempted. Perhaps the import scripts could parse the file to check for such errors before a load is attempted. Stephen
Stephen Hedges wrote:
Paul -
I'm not getting any warnings when I make this change to bulkmarcimport.pl and run it. Just to be sure, I also commented out the line "$batch-
warnings_off();" in dumpmarc.pl and ran the file through that script -- no
warnings.
I've attached a copy of the file of MARC records ("marctestout"). Our mapping is at http://koha.athenscounty.lib.oh.us:8080
http://bugs.koha.org/cgi- bin/bugzilla/show_bug.cgi?id=516
------- Additional Comments From paul.poulain@free.fr 2003-07-03 02:39 ------- strange... it works fine for me (with UNIMARC files). Could you check that your MARC record is correct. just add : warn "==>".$record->as_formatted(); after :$record = MARC::File::USMARC::decode(char_decode($record-
as_usmarc(),$char_encoding));
OK, I understand the problem... 952 it the tab you use for ITEMS. So it's stored with items, NOT with biblioitems. Thus, it's not transmitted to the correct function.
what's the solution ? the only one is to move itemtype from 952 to something else. Why is it coded like this ? it's an historical behaviour (from 1.2.x) biblioitems is supposed to store infos common to some/many items, and there can be more than one for a given biblio. It's something that doesn't exist in MARC, who only has 2 different levels (biblio & item)
One sample : * biblio => 1- the 2 towers, from tolkien * biblioitem => 1- edited in 1957, by "greatbook edition". 2- edited in 2002, by "pocket edition". * items => 1- barcode 100001, biblioitem 2 2- barcode 100002, biblioitem 1 3- barcode 100003, biblioitem 2
HTH. Note that an error should be reporter in checkmarc, i'll investigate why it does not.
PS : i copy this mail to bugzilla and mark #516 as invalid.
-- Paul POULAIN Consultant indépendant en logiciels libres responsable francophone de koha (SIGB libre http://www.koha-fr.org)
Stephen Hedges wrote:
The "only solution" is not a solution at all. The only place the itemtype exists in a MARC record is in the holdings tag (be that 952, 995, whatever), so Koha _must_ be able to move that info into the biblioitems table. The user should not have to invent a workaround!!
Just a general note -- DATA IMPORT PROBLEMS ARE CRITICAL!! The libraries you want to see using Koha are already going to have their data in a commercial ILS, and Koha must be able to smoothly incorporate that data into its tables.
I appreciate the beauty of the Koha database structure, but the import scripts cannot be limited by that structure. Since the itemtype info is located in the MARC holdings tag, then Koha has to retrieve that info from the MARC tag and put it in the appropriate table. If items attached to a MARC record have different itemtypes, that is a MARC cataloguing error and should generate an error message when an import is attempted. Perhaps the import scripts could parse the file to check for such errors before a load is attempted.
Stephen
This is not a problem of the koha IMPORT tools. this is a problem with the DB structure, and internal behaviour. I think this is a MARC problem in fact. Let me explain : if you have some really differents items (various itemtypes, various edition date, various editors), do you have 1 MARC record ? I think that no. You have one MARC record for each different kind of "material" you have. I suspect that MARC21 first idea was to have 1 record for each item (but maybe i'm wrong, it's just a feeling when seeing 050-060 & 760-780 tags). In the default DB, the field itemtypes is mapped to 300$f. If i read all 300 subfields, i see that it speaks of "item physical description", that's what biblioitems is supposed to speak of. How to solve this problem if you have various itemtypes on the same MARC record ? i think the solution is to duplicate your records. Sorry for this bad news, but i don't think it's a bug in Koha. -- Paul POULAIN Consultant indépendant en logiciels libres responsable francophone de koha (SIGB libre http://www.koha-fr.org)
Stephen Hedges wrote:
The "only solution" is not a solution at all. The only place the itemtype exists in a MARC record is in the holdings tag (be that 952, 995, whatever), so
Note that there is the problem with dewey too (952k), that should be 082a. It's wrong in default MARC21. -- Paul POULAIN Consultant indépendant en logiciels libres responsable francophone de koha (SIGB libre http://www.koha-fr.org)
participants (2)
-
paul POULAIN -
Stephen Hedges