In T354566: [Data Quality] [SPIKE] Can we migrate the anomaly detection job to DeeQu checks we developed a PoC to report metrics at row level.
This was useful in the context of anomaly detection, where we wanted to report metrics at country granularity.
We have an increasing need of this kind of instrumentation across pipeline (actor signature, mw history).
Deequ offers a RowLevelSchemaValidator for row level filtering that could help support these use cases.
We need to add alert and metrics constructors to wrap that API in our iceberg data quality tables exporters.