Check a measurements.json against the release it was built from

One row per check with an ok flag, so a release chunk can print the table whether or not it passes. The counts are re-measured against obs_env when con is given (the release's own connection, or a promoted release read back); without it the arithmetic that lives inside the record is still checked.

Usage

check_measurements_catalog(
  record,
  con = NULL,
  dataset_record = NULL,
  measurement_type = NULL
)

Arguments

record

from build_measurements_catalog() (or a path or URL)

con

a DBI connection holding obs_env; NULL to check only what the record can prove about itself

dataset_record

the datasets.json record (or its path) whose datasets[] every dataset_key must appear in; NULL to skip that check

measurement_type

the registry every series[].measurement_type must be in (a data frame, or NULL to skip)

Value

A tibble: check, level, ok, expected, observed, detail.

See also

assert_measurements_catalog(), measurements_catalog_checks()