dr_quality_rows()

Explicitly reads row data for local repair. Rows are never added to catalog metadata or portable quality reports. Formula predicates are re-evaluated on the retained candidate, so keep their captured constants unchanged. Required values and duplicate keys are also supported. Aggregate and custom agent checks do not necessarily identify rows and are rejected with an explanation. Failed lake candidates remain available until cleanup removes them.

Usage

dr_quality_rows(x, rule = NULL, contract = NULL, limit = 100)

Arguments

x: A blocked run result, or a table to diagnose explicitly.
rule: Check name from dr_quality_report(), or a one-sided row formula. Omit it when the result has one failed check or a failed lookup. When several checks fail, the error lists the names to choose from.
contract: Contract when supplying a table and a named check.
limit: Maximum returned rows; Inf explicitly requests all rows.

Value

A tibble containing the affected rows, bounded by limit.

Examples

orders <- dr_product("orders", data.frame(amount = c(10, -2))) |>
  dr_add_quality(list(positive = ~ amount >= 0))
failed <- dr_run(write = FALSE, orders, stop_on_failure = FALSE)
dr_quality_rows(failed, "positive")

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