Occurrences without any taxon

[gbif-portal]

Occurrences without any taxon

5 something observed in Togo

It is worth very little. This occurrences do not even have a verbatim name. Might it be worth adding a new type of issue for cases where no taxon is provided (not just unmatched, but not provided).

---

fbitem-occurrence1413023764
Reported by: @MortenHofft
System: Chrome 55.0.2883 / Mac OS X 10.10.5
Referer: https://demo.gbif.org/occurrence/1413023764
Window size: width 1440 - height 776
API log
Site log
datasetKey: f83b0c09-4ef6-4c6c-9868-019709ba4820
publishingOrgKey: 50e959f4-fec1-44b7-9e03-04fab5cfd3c3

[MortenHofft]

there are 703,245,140 in total

there are 700,139,947 results when searching for all kingdoms

310,789 results of these has kingdom without taxon match

3,327,811 is without taxon match in total

that raises questions
Sometimes TAXON_MATCH_NONE maps to incertae sedis and other times not. why not always use that kingdom key when there is no match?

i would think that
total = all_kingdoms - all_kingdoms_without_taxonmatch + wihtout_taxonmatch_total
but there is a difference on 88171

That indicates that we have 90k occurrences that do not have a kingdom, nor incertae sedis and isn't unmatched either. That seem like something i would like to know as a publisher, but how to find those records?

no taxon match includes occurrences with no verbatim taxon provided as can be seen here

[mdoering]

See also recent commit gbif/occurrence@bf744aa to fix very related issue http://dev.gbif.org/issues/browse/POR-2150

[MortenHofft]

Thank you. That largely solves the issue.

I still find it worth a discussion whether no-taxon-match is the correct remark for occurrences missing a name altogether. Currently we do not mark missing dates as invalid-date for example. Wouldn't a NO_TAXON_PROVIDED info remark be a more informative and useful flag?

[cgendreau]

moved discussion to gbif/gbif-api#11