While a long analytical query is running, a pipeline commits new rows to the same table. What does the table format guarantee for the running query?
Select an answer to reveal the explanation.
Short Explanation
You get snapshot isolation: the running query keeps reading the version it started with. That is why a report running at the same time as the nightly load still adds up.
Full Explanation
The correct answer is the option stating that the running query continues against the consistent snapshot it started from, so its results are not corrupted by the concurrent write. Readers operate against the snapshot current when they began. A commit creates a new snapshot without disturbing the one in use, so a long-running query produces internally consistent results. The option saying the query fails and must be resubmitted once the write completes is incorrect because concurrent writes do not abort readers; avoiding that is a central benefit of the design. The option saying the query automatically incorporates the new rows partway through execution is incorrect because partway incorporation would make results inconsistent - some of the query reading before the write and some after. The option saying the query blocks the write until the query finishes is incorrect because readers do not block writers here; that would make a busy analytical workload starve the pipelines. For the exam, describe reader behaviour as snapshot-consistent rather than blocking or aborting.