library filters ordered by item library code (not quite alphabetical)

[theLinkResolver]

https://www.library.ualberta.ca/symphony?q=

MacEwan comes before Keyano, presumably because the list is sorting by the item codes (e.g. GR_MACEWAN). There are a few other examples in the list for AHS, U of A, Concordia, etc.

[ghost]

Since only the codes are present in the data, can currently only sort by those (the actual library name is dynamically mapped in the UI). I can probably fix this by creating a new MARC mapping file in the application...

[ghost]

Solution for this is the same as for the institutional filter ( #867). I'm afraid this will break something, though...

[ghost]

It's tricky to untangle this. Since it isn't a major issue, I'm going to move this to the summer project milestone.

[ghost]

Done when:

• [ ] Library facet sorts alphabetically

[ghost]

So, we must have made some changes to the codes or the labeling, because this no longer seems to be an issue. Grant Macewan comes before Keyano, which is correct. We still have a problem with the codes showing rather than the labels (because we don't have a good process for staying up to date), but this problem at least seems to have fixed itself. @theLinkResolver If you can find an example of where this still broken, please let me know.

[theLinkResolver]

@redlibrarian https://www.library.ualberta.ca/advanced - pop open the Library facets

Agreed it does seem to be much better than it was, but there are still a couple things:

• Alberta Gov't Internet before Capital Boulevard
• Concordia University before Concordia Seminary

Also the UA library sorting is notable, and perhaps we need to be sure why there are sorted this way and if it's the desired behaviour -

• Winspear files under Business
• Rutherford files under Humanities
• Cameron files under Science
• Bruce Peel files under Special Collections

This is probably due to sorting by library policy, i.e. UABUSINESS, UAHSS, UASCITECH, UASPCOLL

Actually I think that's why we see the behaviours above, as well - AGINTERNET before AGL_CAP and CONCORDIA before CONCORDSEM. The sorting within institutions is by library policy, and within the overall list is alphabetical by text. Does that make sense?

[ghost]

So, we use the library code (I guess what you're calling policy) as the actual piece of data for the facet, and then we just map to the actual full name in the display. So the sort logic operates on the library codes, which does explain the problem we're seeing. I didn't look at the advanced search, but now I see what you mean - because it has the full list (pre-search) of libraries, the problem becomes apparent.

The solution is likely to do what we do with institutions, which is to map the display name to the code during indexing so that the display name and not the code is the actual data-point.

[theLinkResolver]

That all makes sense. And yes, 'library policy' is Sirsi jargon!

[ghost]

@theLinkResolver OK, thanks! I've written this up as #1178 for the dev team, so I'm going to close this issue.