`occ_search()` with `recordedByID` misses records where the field contains multiple IDs

[stevenpbachman]

Not sure if this is an rgbif or GBIF issue. occ_search() with recordedByID seems to miss records where the field contains multiple IDs

occ_search(recordedByID = "https://www.wikidata.org/wiki/Q6525437") returns only records where that URI is the only value in the recordedByID field. Records where the field contains multiple identifiers (e.g. a Wikidata URI, a Harvard UUID, and a VIAF URI concatenated together) seem silently excluded from results, even though the queried ID is present.

Returns ~1 record:
occ_search(recordedByID = "https://www.wikidata.org/wiki/Q6525437", limit = 20)

But this record (GBIF key 3028984454) has Q6525437 in its recordedByID field alongside other IDs, and is NOT returned by the search occ_get(key = 3028984454, fields = "all")

[jhnwllr]

Hi @stevenpbachman, Thanks for writing.

I don't think I understand what the expected behavior is supposed to be.

When you search recordedById = https://www.wikidata.org/wiki/Q6525437, you get back 1 record.

https://api.gbif.org/v1/occurrence/search?recordedByID=https%3A%2F%2Fwww.wikidata.org%2Fwiki%2FQ6525437
https://www.gbif.org/occurrence/search?recorded_by_id=https:~2F~2Fwww.wikidata.org~2Fwiki~2FQ6525437&advanced=1&occurrence_status=present

Why would you expect more records? Expecting partial matching?

[stevenpbachman]

thanks @jhnwllr Yes, I was hoping that searching on recordedById = https://www.wikidata.org/wiki/Q6525437 would return other records like: https://www.gbif.org/occurrence/3028984454, which also has the same recordedById Q6525437, although this is concatenated with other IDs. So, yes, I guess this would require partial matching, otherwise I'm not sure how I would pull all records that matched Q6525437.

[jhnwllr]

@stevenpbachman

You might be interested in running a GBIF sql download if you need partial matching. See example below:

library(rgbif)

sql <- "
SELECT
  gbifid,
  datasetkey,
  scientificname,
  eventdate,
  countrycode,
  recordedby,
  recordedbyid
FROM occurrence
WHERE GBIF_StringArrayLike(recordedbyid, '*Q6525437*', FALSE)
"

occ_download_sql(sql)