We would like to add 780$x and 785$x to the following Elasticsearch MARC21 index mappings: - identifier-standard - issn - issn-all (Bug 36727)
I am not sure if that is what every library will expect. These are ISSN of different works and the results in searching might be surprising to some. Of course they can change the mappings locally - might need some thought to decide what is best to offer by default.
We would be okay with adding this just to the issn or issn-all indexes, if you feel that is more appropriate? 780$x and 785$x do represent ISSNs so would be useful for cataloguers to search on in this way: * https://www.loc.gov/marc/bibliographic/bd780.html * https://www.loc.gov/marc/bibliographic/bd785.html
That's actually what gave me some headache. A patron/user will search for a known ISBN, but find totally different records/titles and the sorting could also not be ideal I worry if that fits patron expectations.
I understand. If none of these 3 indexes seems appropriate, would you prefer that we make a different one? issn-serial for example.
I am adding a few names here, having some different perspectives might help.
(In reply to Katrin Fischer from comment #3) > That's actually what gave me some headache. A patron/user will search for a > known ISBN, but find totally different records/titles and the sorting could > also not be ideal I worry if that fits patron expectations. .. ISSN of course.
I definitively agree that it would be nice to have this subfields indexed in ES and to be able to search with their content, but I'm not sure if issn / issn-all is the right place. I agree with Katrin's comment #3. So, maybe what we needed would be a nice name for a new ES search field to put there maybe all standard identifiers from 77X/78X fields ($x, $z). This would enable us to search with the respective search terms in a general keyword search. How about identifier-linking ? Just an idea.
That sounds reasonable to me, but I'll just check with the sponsor that it will work for them :)
We won't be working on this one for now, but happy for anyone else to pick it up
Thanks Janusz for weighing in.
I think the same problem you are fearing, Katrin, is already happening with title search since 780$t and 785$t are indexed in the title. But I agree that an ISSN search, like an ISBN search, is more precise than a title search and should return less results. While looking at the existing mappings, I saw there was one index called related-periodicals. I don't know where/how it is used, but it could be something like that. A new index called related-issn, or even add the 780$x and 785$x to related-periodicals. What is the exact use case Aleisha? They want to use the index drop-down, choose ISSN, type in an ISSN and find all related periodicals? Or is it for an integration with another tool?
Also, maybe unrelated, but I learned of the recently added 023 for Cluster ISSNs. https://www.loc.gov/marc/bibliographic/bd023.html From what I understand, it's an ISSN that represents a number of other ISSNs (?). It's something we'll eventually have to consider.
(In reply to Caroline Cyr La Rose from comment #11) > > What is the exact use case Aleisha? They want to use the index drop-down, > choose ISSN, type in an ISSN and find all related periodicals? Or is it for > an integration with another tool? Originally they had been interested in both the 'standard number' and 'ISSN' advanced search dropdowns also searching the 780$x and 785$x In saying that they have decided this is no longer a priority for them, so if there isn't agreement on the best way to make these tags searchable then we can probably close this report for now.