Organism quantity integer is changed to float in record view (biocache)

[sat01a]

eg https://biocache-test.ala.org.au/occurrences/4c433838-cace-4ab3-bb01-143d44a9f2af

[adam-collins]

Clarification required. There are occurrences without whole number quantities, e.g. https://biocache.ala.org.au/occurrences/search?q=organismQuantity:0.6667

[adam-collins]

@peggynewman what are your thoughts?

[peggynewman]

My thoughts are that it should probably just be displayed as provided, otherwise we are insinuating meaning ourselves.

[adam-collins]

The suggestion is to remove the processed organismQuantity value. This is a Double value stored as a String in SOLR. There is no associated quality assertion when a non-parsable Double value is supplied.

Some options that I think are in line with this request:

• Remove the processed organismQuantity, deviating from the GBIF pipeline process.
• Remove the processed organismQuantity from the GBIF pipeline. This would align GBIF with the example BraunBlanquetScale and similar.
• Alter the processed organismQuantity value stored in the SOLR String field to remove the .0.
• Display only the raw_organismQuantity in on biocache page. This would only ignore the processed organismQuantity on that one page.
• Use the biocache-service field translation to replace all requests of organismQuantity with the raw_organismQuantity. This avoids pipeline changes but could produce a download with the field duplicated.

Of interest may be the unprocessed raw_organismQuantity values https://biocache.ala.org.au/ws/occurrences/search?q=raw_organismQuantity:*&fq=-organismQuantity:*&facet=true&facets=raw_organismQuantity&pageSize=0&flimit=-1

[peggynewman]

The above was from the sampling-event docs for the IPT. This is the DwC definition:

The values that didn't process are a lot of ranges, and those that did are either integers or floats, with quite a few integers being estimates (eg "25000") so to add the precision of a .0 isn't helpful I don't think.

My take is to follow the standard as closely as we can, so option #2 looks best for me. If the GBIF pipeline needs changing then we can do that. What do you think?

[adam-collins]

Moving the issue to https://github.com/gbif/pipelines/issues/922

[adam-collins]

pull request https://github.com/gbif/pipelines/pull/1002

[peggynewman]

1 is no longer 1.0 https://biocache-test.ala.org.au/occurrences/4c433838-cace-4ab3-bb01-143d44a9f2af 0.667 still works https://biocache-test.ala.org.au/occurrences/search?q=organismQuantity:0.6667#tab_recordsView raw_organismQuantity is gone LGTM, tick