CF variable metadata from the measurement_type registry

Turns the canonical measurement registry into the per-variable lookup the netCDF writers use: a variable needs one units, one long_name and optionally one standard_name. This is the widening half of publishing — the database stores every quantity in a single measurement_value column, so the units live in the registry rather than on the value.

Usage

measurement_var_meta(mt)

Arguments

mt

data.frame from read_measurement_type(). Only measurement_type is required; units, description, standard_name, is_canonical, valid_min and valid_max are used when present.

Value

Named list keyed by measurement_type, each element list(units, long_name, standard_name, canonical, valid_min, valid_max). An empty registry cell becomes "" for units (never the string "NA") and falls back to the type name for long_name, because a CF variable with long_name = "NA" is worse than one with no long_name.

Examples

mt <- data.frame(measurement_type = "temperature_ave", units = "degree_C",
                 description = "average temperature", stringsAsFactors = FALSE)
measurement_var_meta(mt)$temperature_ave$units
#> [1] "degree_C"