The Performance Recording indicates that "Executing Query" is the longest-running event, meaning the query sent to the data source is the primary performance bottleneck. A Relative Date filter on a continuous date field (Created Date) over a five-year span forces the database to evaluate a wide range of values. Replacing this with a filter on a discrete, calculated field like YEAR([Created Date]) significantly improves performance. This change simplifies the query's WHERE clause, allowing the database to filter on a small set of integer values (the years) instead of a large, continuous range of dates, which is a much more efficient operation.

A. Replacing the filter type addresses the performance of rendering the filter control itself, not the "Executing Query" event, which is the identified bottleneck.

C. TOTAL() is a table calculation computed after the main query returns data. Modifying it will not improve the initial query execution time.

D. Replacing a simple aggregation (SUM) with a table calculation (WINDOWSUM) would shift computation from the database to Tableau and likely worsen performance, not improve the query time.

1. Tableau Desktop and Web Authoring Help

"Optimize Workbook Performance":

Section: "Filter efficiently" -> "Filter by creating a discrete field"

Content: "When you filter on a continuous date

you can select a range of dates

which requires a database query. But if you filter on a discrete date

you can select specific dates

which is a more efficient query." This directly supports the principle behind replacing a continuous date filter with a discrete one (like YEAR) to improve query performance.

2. Tableau Whitepaper

"Designing Efficient Workbooks":

Section: "Filters" (typically found under "Data Source and Query Performance")

Content: The document emphasizes that filters on discrete fields with lower cardinality (like years) tend to perform better than filters on continuous fields with high cardinality (like a full date range). This is because the resulting query is simpler and can be processed more quickly by the underlying data engine (Hyper

in this case).

3. Tableau Desktop and Web Authoring Help

"Table Calculations: Addressing and Partitioning":

Section: "Performance"

Content: This documentation explains that table calculations like TOTAL() and WINDOWSUM() operate on the table of aggregated data that has been returned to Tableau. Since they are post-query calculations

optimizing them does not resolve a bottleneck in the "Executing Query" phase. This confirms why options C and D are incorrect.