When employing window functions versus table functions in Snowflake, how do they differ in their application and output?

Question

Options

• ATable functions return tables as their output
• BWindow functions operate on entire datasets and generate tables as output
• CWindow functions provide aggregate results and modify table structures directly
• DTable functions process data within specified partitions or frames

• A
  3% (1)
• B
  10% (3)
• C
  13% (4)
• D
  74% (23)

Explanation

There appears to be an error in this question's answer key. Option D states that table functions process data within partitions or frames - but that is actually the defining behavior of window functions, not table functions. The answer key has them swapped.

Here's what's actually correct:

• A is TRUE about table functions: they do return tabular results (a set of rows), used via TABLE() or LATERAL in a FROM clause.
• D is TRUE about window functions, not table functions: OVER (PARTITION BY ... ORDER BY ... ROWS BETWEEN ...) defines the partition/frame a window function operates on.
• B is wrong because window functions don't generate tables as output - they add a computed column per row while preserving the original row count.
• C is wrong because window functions never modify table structures; they're purely read-only computations in a SELECT.

Memory tip: Think "window = frame around rows" (partitions/frames) and "table function = returns a table". The word in each name tells you exactly what it does.

If this is from an official practice exam, I'd flag D as a likely typo - the correct answer describing table functions should be A.

No community discussion yet for this question.