Open tpendragon opened 9 years ago
So for example, if you wanted both Author and Creator to show up in the same field, you'd have to process the solr document before it went into solr to add a combined_author_and_creator field. If you changed those settings you'd have to reindex everything.
Well, or atomically update everything, but both are expensive operations.
I wonder if we can do some of this automatically. For instance:
Let's say we want to combine authors, photographers, contributors, arrangers, artists, creators etc all under one facet.
Those predicates are
Author: http://id.loc.gov/vocabulary/relators/aut
Photographer: http://id.loc.gov/vocabulary/relators/pht
Contributor: http://purl.org/dc/elements/1.1/contributor
Creator: http://purl.org/dc/elements/1.1/creator
If you get the triples for author/photographer predicates, they're sub-properties of Contributor (http://purl.org/dc/elements/1.1/contributor)
So if you were to automatically populate all predicates into a solr field for their lowest level subproperties, then you'd end up with author/photographer/etc all automatically grouped together in a place that you can facet.
The problem only comes with DCE.creator - creator's not a subproperty of contributor, and they're not related in any way (which is...odd to me). Maybe this is a metadata problem and we should be using contributor? If not, then automatic grouping only gets you so far.
I just pushed something up, but I'm still looking at Solr hierarchical facet info; in fact it looks like there is a project that uses Solr pivots with Blacklight https://groups.google.com/forum/#!topic/blacklight-development/pLSOhkyd2GI so I'd like to spend a few more days on this. On the other hand, the docs do mention that there is a performance hit.
Oregon Digital 1 ticket: OregonDigital/oregondigital#269
So I was right - the only way to do this in Solr is to generate these fields.