Hello Peter,
I can't say we've had anything like this reported to us.
FWIW, I grabbed the latest bulk downloads of Names, and the preceding three downloads, which makes the oldest one from 2016, and the labels for "Yuan, Longping, 1930-" all include a language tag. (The 2016 download seemed to capitalize the EN.) That particular selection spans two different systems/ways the bulk downloads were generated.
I did not check LCSH but I suspect I would get the same results, if for no other reason because the same code produced the RDF for both Subjects and Names and the same code produces the bulk downloads.
Is it possible that you stripped the language tags before previously loading the data? Is it possible that two different versions of Fuseki were used, one that had a bug related to language tags or different support, which may or may not be a factor?
Not sure this was terribly helpful.
All the best,
Kevin
--
Kevin Ford
Library of Congress
Washington, DC
-----Original Message-----
From: LC Linked Data Service Discussion List <[log in to unmask]> On Behalf Of Peter M. Green
Sent: Monday, February 04, 2019 4:40 PM
To: [log in to unmask]
Subject: [ID.LOC.GOV] Fewer results with latest dumps
After some massaging (using sed), I managed to upload the lcnaf and lcsaf dumps into Fuseki and have been puzzling over why queries that retrieved URIs for names and subjects from the previous dumps, and which do return data from the label service now, aren't returning any results.
Examples:
lcnaf: SELECT ?s WHERE { ?s ?p "Yuan, Longping, 1930-"@en . }
lcsaf (same query): Beat generation
Anyone else seeing this?
Thank you,
Peter
|