The first task in a Lakeflow Job, named qc, ends with `dbutils.jobs.taskValues.set(key = "bad_records", value = bad_count)`. If bad_records is greater than zero, the quarantine notebook must run; otherwise the publish notebook must run. How should the data engineer implement this branching?

1. A. Add a For each task whose inputs are the values true and false, and let a wrapper notebook decide which path to call 💬 A
2. B. Add an If/else task with {{tasks.qc.values.bad_records}} > 0, with quarantine on true and publish on false ✓ B
3. C. Add quarantine with a run-if condition of All failed and publish with All succeeded, both depending on qc 💬 C
4. D. Pass bad_records as a job parameter to both notebooks and let each notebook exit early when it should not run 💬 D

Reveal answer

Correct answer: Add an If/else task with {{tasks.qc.values.bad_records}} > 0, with quarantine on true and publish on false

Reference: https://docs.databricks.com/aws/en/jobs/tasks/if-else