Data-Engineer-Associate - Working with Lakeflow Jobs (16% of the exam) - Section 4.4

Choose between time-based and data-driven triggers based on data availability and pipeline dependencies.

Given how and when upstream data arrives, choose a time-based schedule or a data-driven trigger, and justify the choice against wasted runs, latency and missed data. Recognise when a continuous job is the better answer.

time-based triggerdata-driven triggerdata availabilitycontinuous jobsupstream dependencies

Practice question for this objective

Free sampleWorking with Lakeflow Jobshard

A gold Lakeflow Job aggregates a silver Unity Catalog table that three separate upstream jobs write to at different times of the morning, and their finishing times drift by up to ninety minutes. The gold job currently runs on a cron schedule chosen to sit after the slowest upstream job, so it usually reads stale data on fast mornings and occasionally reads a half loaded table on slow ones. The team cannot change the three upstream jobs. Which trigger should the gold job use?

  • AA file arrival trigger on the cloud storage path that holds the silver table's underlying data files in the metastore.
  • BA continuous trigger, so the gold aggregation is always rerunning and therefore always reflects the latest state of the silver table.
  • CA scheduled trigger set to a much shorter interval, with a first task that exits early when the silver table shows no new version.
  • DA table update trigger on the silver table, so the gold job starts a run when a new commit is made to that table. Correct
Use a table update trigger when a downstream job must react to upstream writes it cannot observe through the upstream job definitions. The gold job needs a signal for upstream completion, and the only shared artefact is the silver table. A table update trigger fires on new commits to that Unity Catalog table, so the aggregate runs once the data actually changes rather than at a clock time picked to cover the worst case upstream finish.

Why A is wrong: It sounds equivalent because a Delta write does add files, but the file arrival trigger is meant for a landing location of raw files and would fire on intermediate file activity rather than on a committed table version.

Why B is wrong: Continuous execution would keep the aggregate fresh, but it burns compute all day for a source updated three times a morning, and the team asked to react to upstream completion rather than to run without pause.

Why C is wrong: This is a workable but cause blind fix: it rebuilds trigger logic by hand inside the job and still starts many runs whose only purpose is to discover that nothing has changed.

Why D is correct: A table update trigger watches for commits to the named Unity Catalog table, which is the only signal the gold job can observe that reflects the upstream writes without touching the upstream jobs.

See more Data-Engineer-Associate practice questions, answers explained.

Exam traps in Working with Lakeflow Jobs

Answers that look right on this material and are not. Each one is a distractor from a different question in the Data-Engineer-Associate bank for this domain.

  • Both belong on data driven triggers, because a trigger that reacts to arrival is strictly more current than one that reacts to the clock, and the regulatory extract can carry a task that writes an empty result if no data arrived.

    Why it is wrong: The reasoning about currency holds for the partner feed, which is why the sweep is tempting. It is wrong because a data driven trigger has nothing to react to on a day with no source change, so an obligation to produce something daily would go unmet.

  • A scheduled trigger every minute with concurrent runs raised, so that a new run starts even while the previous one is still processing events.

    Why it is wrong: Allowing concurrency removes the skipped runs, which is why it looks like the fix, but it starts overlapping streaming queries against the same checkpoint and source, which is a contention problem rather than a latency fix.

  • A file arrival trigger on the downstream job, pointed at the cloud storage path that holds the silver table's underlying data files.

    Why it is wrong: It appeals because a Delta write does produce new files, but a file arrival trigger is meant for a landing location of raw files and would react to any file the table writes, including compaction output, rather than to a committed table version.

Examworthy is not affiliated with or endorsed by Databricks. Original, blueprint-aligned practice material only.