Build the ExtendedMeasurementOrFact extension for one dataset

Three grains, all with measurementTypeID / measurementUnitID from measurement_type.csv's nerc_p01 / units_nerc_p06 — empty where the registry states no exact concept, never invented:

Usage

dwc_emof(
  con,
  dataset_key,
  occurrence = NULL,
  measurement_type = NULL,
  env = TRUE
)

Arguments

con

a DBI connection to the release

dataset_key

the dataset

occurrence

the frame from dwc_occurrence() — needed to resolve occurrenceID; without it the obs_attribute grain is skipped

measurement_type

the registry from read_measurement_type(), or NULL

env

include obs_env rows on the dataset's own events

Value

A data frame of eMoF rows, all-NA columns dropped.

Details

• event — sample_measurement rows for the dataset, minus volume_sampled (which is the event's sampleSizeValue, not a repeat measurement).

• occurrence — obs_attribute rows, joined to their occurrence on (sample_key, taxon_key, life_stage). Which of a bin's two numbers is the measurementValue is decided by the REGISTRY, not per dataset: a type with a physical units (body_length mm, carapace_length mm) puts bin_value in measurementValue and the bin's count in measurementRemarks; a type with no unit (stage, behavior) is a categorical bin, so the count is the value and the bin_label is the remark.

• event, environmental — obs_env rows sitting on one of this dataset's own events (sample_key). Empty for a dataset that measures no environment itself.

See also

dwc_event(), dwc_occurrence()