Free DP-700 flashcards

In Microsoft Fabric, which Spark workspace setting lets notebook sessions start within seconds without permanently reserving compute?

Keeping the starter pool enabled. The starter pool holds pre-warmed, pre-allocated nodes attached to the capacity, so a session binds to a ready node in seconds instead of paying cold-start provisioning time, and the nodes are released when idle rather than reserved continuously.

How do you move an existing Microsoft Fabric workspace onto a purchased F-series capacity without recreating its items?

Reassign the workspace to the target capacity in its settings. Capacity assignment is a workspace setting, not baked into each item, so reassigning repoints every Lakehouse, notebook, and Dataflows Gen2 to the production compute in place at once.

Common misconceptionThat moving a Fabric workspace to a new capacity means exporting every item and rebuilding it in a fresh workspace. Reassigning the capacity in workspace settings moves everything in place instead.

How do you apply a daily delta of new and changed customers to a current-state silver Delta table without keeping prior versions?

Use a MERGE keyed on the business key (such as CustomerId): matched rows are updated in place and unmatched rows are inserted. This Type 1 upsert handles both new and changed customers atomically while leaving unchanged customers untouched. Append would create duplicate keys and overwrite would delete customers absent from the day's delta.

Common misconceptionThat overwriting the whole table with a daily delta keeps it current. The delta lacks unchanged customers, so overwriting deletes every customer absent from that day's extract.

How should you incrementally load an append-only source with an ever-increasing timestamp into a bronze Delta table while keeping overhead low?

Store the highest timestamp value loaded so far as a high-water mark, then each run copy only source rows whose timestamp exceeds that stored value and update the mark. Because rows are only inserted and the timestamp always increases, the previous maximum cleanly separates old from new rows, so you load exactly the new data without reading the whole table or enabling source-side change tracking.

Common misconceptionThat truncating and reloading an append-only source nightly is low overhead. It rereads the whole source every run, the exact cost a watermark avoids.

In Microsoft Fabric, where do you see recent runs of pipelines, Dataflows Gen2, and notebooks together in one status view?

The Monitoring hub is the central place that aggregates run activity across item types into a single filterable list showing each run's status and start time. This lets you spot overnight failures across many Data Factory pipelines, Dataflows Gen2, and Spark notebooks at once, without opening each item's own history.

How can a Fabric Copy activity recover automatically from transient source failures without a separate wrapper pipeline?

Set the Copy activity's retry count to a small positive value with a retry interval. The activity then re-attempts the copy automatically after a transient failure, which clears brief connection resets that succeed on rerun, with no extra pipeline to build.

Common misconceptionThat raising a Copy activity's timeout fixes a transient connection reset. Timeout only limits how long an attempt may run; it does not re-establish a dropped connection, so retry count and interval are needed instead.