How to Write SQL Window Functions with AI Quiz

Answers and explanations appear as you take the quiz.

1. For ordered purchases of 5, 7 and 4, which row-by-row running totals match the guide's frame?
   • 5, 7 and 4
   • 5, 12 and 16
   • 16, 16 and 16
   • 5, 14 and 12
2. For descending scores 100, 100 and 90, which sequence does RANK produce?
   • 1, 2 and 3
   • 1, 1 and 2
   • 2, 2 and 3
   • 1, 1 and 3
3. Which function gives tied scores 100, 100 and 90 the ranks 1, 1 and 2?
   • DENSE_RANK
   • ROW_NUMBER
   • RANK
   • LAG
4. A latest-event query uses ROW_NUMBER ordered only by a timestamp shared by two events. What is needed for a predictable choice between them?
   • Remove the account partition
   • Replace the final result with an unqualified DISTINCT
   • Add an appropriate tie-breaker to the window ordering
   • Assume insertion order resolves every tie
5. Why can LAG of daily sales fail to represent yesterday's sales?
   • LAG always refers to the first row of the entire table
   • The previous recorded row may be separated by missing dates
   • LAG automatically groups all dates in one month
   • The function replaces missing dates with zero
6. An ORDER BY appears inside OVER. What is still needed if the final displayed result must have a specified order?
   • Only a PARTITION BY with the same column
   • Only a ROWS frame using the same ordering column
   • No further clause because window ordering guarantees display order
   • A final ORDER BY for the query result
7. Why should a tie-breaker not be added indiscriminately to RANK when equal scores should share a rank?
   • It can change which rows are peers and therefore change tied ranks
   • It leaves peer membership unchanged because only the first ordering column matters
   • It changes only the final display order and never the window calculation
   • It preserves equal-score ranks while automatically making their row numbers predictable
8. In PostgreSQL, how can a query filter rows using a window value calculated in an inner SELECT?
   • Use that value in the same SELECT's WHERE without another query level
   • Replace OVER with GROUP BY in every case
   • Filter in an outer query or common table expression
   • Assume every database supports an identical QUALIFY clause