Check that a DGF's rules still fit the data

For every variable carrying a `raw_to_stable_transform` or `desired_data_import_rule`, applies the rule to the matching column of `data` and measures the share of *present* values it turns into `NA` — the same signal the rule's `is_*` guard would give. A high miss rate means the rule no longer matches the data (e.g. you cleaned a date column, so the old `read_mmddyyyy` no longer applies).

Usage

check_dgf_rules(dgf, data, threshold = 0.2)

Arguments

dgf

A DGF data frame, or a path to a DGF `.xlsx`.

data

The data to test against: a data frame, or a path to a `.csv`.

threshold

Miss rate above which a rule is flagged as not fitting. Defaults to `0.2`.

Value

A data frame with one row per (variable, rule): `var_name`, `stage` (`"reformat"` or `"import"`), `rule`, `miss_rate`, and `fits` (logical). Empty when the DGF has no rules. Rows with `fits == FALSE` are the ones to review.

See also

[stabilize_data()], [update_dgf()]

Examples

if (FALSE) { # \dontrun{
bad <- subset(check_dgf_rules(dgf, raw_data), !fits)
} # }