The MARC leader position 17 stores a record's encoding level -- # indicates a fully encoded and verified record, with numbers 1-8 indicating increasingly un-encoded and un-verified status. It would be useful to create the option on batch import of only overlaying if the incoming record has a lower (better) encoding level than the existing record. I could see this as either a new option in the "Action if matching record found:" dropdown when staging for import OR as a new variety of match check in the matching rule.
Hmm that's interesting. I suppose other uses could be a match check on say the 005 to make sure that the incoming record's value is '>=' to the existing record.
So maybe a broader move to allow match checks to use a comparison other than equals? So we'd set a target subfield in the incoming record, a target subfield in the existing record, and a comparison operator? But the values on the encoding level aren't strictly numeric. "#" is the "best" value, followed by 1-8, then "u" for unknown and "z" for not applicable (https://www.loc.gov/marc/bibliographic/bdleader.html). We could hardcode that hierarchy for encoding level, but it seems likely other bits of MARC have comparably idiosyncratic sets of values. Should the match check setup include a mechanism for telling Koha an order of preference for possible values?
(In reply to Andrew Fuerste-Henry from comment #2) > So maybe a broader move to allow match checks to use a comparison other than > equals? So we'd set a target subfield in the incoming record, a target > subfield in the existing record, and a comparison operator? > I'd say a broader move to allow match points to use comparisons other than equals. (LDR and control fields don't have subfields.) > But the values on the encoding level aren't strictly numeric. "#" is the > "best" value, followed by 1-8, then "u" for unknown and "z" for not > applicable (https://www.loc.gov/marc/bibliographic/bdleader.html). We could > hardcode that hierarchy for encoding level, but it seems likely other bits > of MARC have comparably idiosyncratic sets of values. Should the match check > setup include a mechanism for telling Koha an order of preference for > possible values? Mmm that's an interesting wrinkle. That's a lot more complicated. Perhaps we should look at plugins for record matching rules.
(In reply to David Cook from comment #3) > (In reply to Andrew Fuerste-Henry from comment #2) > > So maybe a broader move to allow match checks to use a comparison other than > > equals? So we'd set a target subfield in the incoming record, a target > > subfield in the existing record, and a comparison operator? > > > > I'd say a broader move to allow match points to use comparisons other than > equals. (LDR and control fields don't have subfields.) > I'm just reviewing my old bug reports and what do you know. I had this idea 5 years ago heh. Bug 15536.