試験Analytics-Con-301 トピック1 問題100 スレッド

Comprehensive and Detailed Explanation From Exact Extract:
The key detail of the question:
"filter out particular categories, but wants the percent of total calculation to remain steady even as they filter items in or out." This means the percent of total must ignore filters.
Table calculations always operate after filters, except table calc filters like "Filter on Table Calculation," and after dimension filters, so filtering categories directly will change the denominator.
Tableau's documented solution for "percent of total that does not change with filtering" is:
# Use a FIXED LOD to define the stable denominator
A FIXED LOD expression "freezes" the aggregation level and is unaffected by dimension filters unless explicitly added to context.
This allows the consultant to compute:
{ FIXED : SUM([Sales]) }
or
{ FIXED [Category] : SUM([Sales]) }
Then percent of total becomes:
SUM([Sales]) / { FIXED : SUM([Sales]) }
The FIXED LOD stores the total before filters are applied, ensuring the percent remains steady.
This is exactly what Tableau documentation explains under:
* Level of Detail Expressions
* LODs and Order of Operations
* Using LODs to create filter-independent calculations
Thus, D is correct.
Why the other answers are wrong:
# A. Context Filter
Context filters run before FIXED LODs but after raw data.
If Category is put into context, LOD totals would be reduced.
Table calculation totals still change because table calcs run near the bottom of the pipeline.
# B. Data Source Filter
Data source filters remove rows before all table calculations and LODs.
This would make the percent of total incorrect, because filtered-out categories would physically be gone.
# C. Aggregate Expression
An aggregate field alone does not solve the issue because it still respects dimension filters.