Apply a PII field policy to a data frame — bs_apply_field

Filters or redacts columns based on a YAML-driven field policy. This is the same logic the MCP server uses to strip PII before data reaches the AI model.

Usage

bs_apply_field_policy(df, dataset_name, policy = NULL)

Arguments

df: A data frame (typically from bs_get_dataset()).
dataset_name: Character string identifying the BDS dataset.
policy: A named list representing the field policy (as returned by yaml::read_yaml()). If NULL (the default), loads the bundled field_policy.yml shipped with the package.

Value

The input data frame with the field policy applied.

Details

The policy supports three modes per dataset:

allow: Only the listed fields are kept; all others are dropped.
redact: The listed fields have their values replaced with "[REDACTED]"; all other fields pass through.
all: All columns pass through unchanged.

If dataset_name is not found in the policy, the data frame is returned unchanged.

Examples

df <- data.frame(
  UserId = 1L, FirstName = "Jane", Organization = "Org1",
  stringsAsFactors = FALSE
)
bs_apply_field_policy(df, "Users")
#> Error in bs_apply_field_policy(df, "Users"): could not find function "bs_apply_field_policy"