https://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=34482 --- Comment #12 from Martin Renvoize (ashimema) <martin.renvoize@openfifth.co.uk> --- Hi all, While working on bug 35117 I went through the bugs about how Koha links component parts, volumes and sets to their parent records when UseControlNumber is on (001, 003 and the $w of 773 and the 8xx fields), and compared the current behaviour with the MARC 21 definition of $w, "(OrgCode)control number". These are the points I could confirm on a Zebra instance or in the code: - The index behind the volumes search (rcn) covers $w of 770 to 787 and 800/810/811/830, so a set's volumes also include records linked from 776, 780 or 785 (bugs 35330 and 39287). - With a host that has a 001 and a 003, a component with a bare $w and no 003 of its own is not found, because the bare branch also requires the component's 003 to match the host's. A $w of the form (003)001 is found either way. - get_volumes_query does not quote its values like get_components_query does, and the 008/21 "monographic series" test does not look at the Leader or 008/06. - The component sort is not applied (a separate bug to follow). I have added what looks like the relevant work as dependencies here, and suggest this order: 1. Bugs 35117 and 41713: no 001 should not search for an undefined value. 2. Bug 39287, then 35330 (and 38252, 42027): more specific indexes for volumes and components. 3. Bugs 27507, 29268 and 21472: matching $w on the (OrgCode)control number form, which is what the spec defines. I would like to write this up as one design bug. 4. Bugs 31608, 43113, 26480 and 26333: re-test once 1 and 2 are in. I will take 1 and help move 2 along. If anyone wants to take 3, or has already started, please say so here. -- You are receiving this mail because: You are watching all bug changes. You are the assignee for the bug.