Sample Objects for Audit

Selects a random sample of objects that meet a concept definition for human review. This is the foundation of drift detection: by regularly auditing samples, you can detect when definitions diverge from reality.

Usage

ont_sample_for_audit(
  concept_id,
  scope,
  version = NULL,
  n = 20,
  where = NULL,
  concept_value = TRUE,
  seed = NULL,
  con = NULL
)

Arguments

concept_id

Character. The concept to sample from.

scope

Character. The scope.

version

Integer. The version. If NULL, uses active version.

n

Integer. Number of objects to sample.

where

Character. Optional SQL WHERE clause to filter the sample population (e.g., only sample from today's cases).

concept_value

Logical. If TRUE (default), only sample objects where the concept evaluates to TRUE. Set to FALSE to sample regardless of concept value, or NA to sample where concept is FALSE.

seed

Integer. Random seed for reproducibility.

con

A DBI connection. If NULL, uses the active connection.

Value

A tibble of sampled objects with their concept values.

Examples

if (FALSE) { # \dontrun{
ont_connect(":memory:")
# ... setup ...

# Sample 20 cases that the system says are "ready for discharge"
sample <- ont_sample_for_audit(
    concept_id = "ready_for_discharge",
    scope = "flow",
    n = 20,
    concept_value = TRUE
)
} # }