There is a sort of cleaning functionnality in Koha when saving a biblio record. If a field contains only a $9 subfield, the whole field is removed. There is a logic there since $9 is generally used in Koha for storing authority record ID. However there may exist fields perfectly valid containing only a $9 subfield. This is a case in Unimarc 010 field.
Created attachment 73676 [details] [review] Bug 20522: Fields with only one $9 subfield are removed There is a sort of cleaning functionnality in Koha when saving a biblio record. If a field contains only a $9 subfield, the whole field is removed. There is a logic there since $9 is generally used in Koha for storing authority record ID. However there may exist fields perfectly valid containing only a $9 subfield. This is a case in Unimarc 010 field. To test in Unimarc: 1. Add a $9 sufield to 010. You can name it "ISBN d'une livraison de publication en série" 2. Create a new biblio record. Fill 010 $a and $9 subfields. Save the record. => Everything is fine: $a et $9 are properly stored. 3. Modify the previous record, and remove $a subield. Save the record. => 010 field is removed 4. Apply the patch. Repeat 2-3 steps. Observe that a 010 field containing only a $9 subfield can be saved properly. To test in MARC21: Replace 010 filed by any other field...
The question is whether this is a correction of a prior mistake or is this the wrong solution for some reason and if so what.
Thanks for the link to bug 5683. My patch equals to revert it. I'm totally unable to evaluate the consequences. I even don't understand in the first place why an empty field (a field with no subfield) or $9 only field would be an issue.
We at Biblibre have the same issue for our libraries wanting 010$9. Note that when creating a new records, AddBiblio is called and it does not contain this code that removes $9. Bug 5683 says : " It seems that running link_bibs_to_authorities.pl can corrupt records in cases where there is a heading with only subfield 9 populated, since the script automatically removes all subfield 9s. " So I understand that link_bibs_to_authorities.pl must handle the case where field has no subfield anymore, not ModBiblio that is also used by cataloguing page. I'd say it must be fixed in LinkBibHeadingsToAuthorities(). PS : we must also revert tests in t/db_dependent/Biblio.t
Wouldn't it be easier/safer to add the UNIMARC exception "010 with $9" instead of completely removing the code?
(In reply to Marcel de Rooy from comment #5) > Wouldn't it be easier/safer to add the UNIMARC exception "010 with $9" > instead of completely removing the code? Or maybe code must check that field is linked to an authority type.
The French Bibliographic Agency (Abes) advocates that the use of $8 is not recommended. If a temporary sub-field for these data must be choosen, the best is to choose a letter, as the use of letters in UNM is free - contrary to the numbers sub-fields. It could be $y. Context : Earlier this year, the PUC made this subfield $8 a generic to all fields of the Unimarc B format. Its meaning is very precise: "Part of the resource concerned by the data recorded in the field" ("Reserved for Materials specified"); it is in practice especially applicable to certain descriptive fields even if it may be also used for identifiers. Do you agree with $y ? Enssib Team
I'm not in favor of this solution. The issue must be solved at its root rather than fixed afterward.