dr_run_history()

Pass evidence = "runs" to dr_run() to keep one JSON record per run. Evidence contains descriptive metadata and quality counts, never input rows, executable definitions, connections, request objects or raw conditions. URL credentials and query strings are removed. Business descriptions and identifiers remain visible: treat the directory as operational metadata.

Usage

dr_run_history(path, product = NULL)

dr_read_run(path, run_id)

dr_incidents(x)

Arguments

path: Evidence directory.
product: Optional product name to filter.
run_id: Run identifier returned by dr_run().
x: A run result, an evidence record or an evidence directory.

Value

dr_run_history() and dr_incidents() return tibbles. dr_read_run() returns a descriptive list, including catalog delivery state.

Details

Records are written by replacing a file in the same directory. Use a single coordinated writer for a directory, including dr_retry_catalogs(). This is a local evidence store, not a distributed transaction log. Writing data and recording evidence are separate operations; evidence errors never imply that published data was rolled back. A successful write followed by a process crash before recording evidence remains a reconciliation responsibility.

Examples

path <- tempfile("runs-")
result <- dr_product("orders") |>
  dr_add_source(data.frame(id = 1:2)) |>
  dr_run(evidence = path)
dr_run_history(path)
dr_read_run(path, result$run_id)$status
dr_incidents(path)
unlink(path, recursive = TRUE)

Documentation source snapshot · View source at d9cb454 ↗
Code examples are documented source examples; they are not executed in this website build.