# Free SCTDA practice questions

> 10 free SCTDA practice questions with the correct answer and an explanation.

Page: https://devopscodelab.com/exams/tableau-data-analyst/free-practice-questions/
Practice exams: https://devopscodelab.com/exams/tableau-data-analyst/

### Question 1 of 10 (choose 2)

An extract built from a 400-million-row fact table is slow to refresh. Analysts only ever look at the last 24 months and do not need row-level detail below the day. Which two extract settings reduce the extract size while still supporting this analysis? (Choose two.)

- A. Add an extract filter on Order Date for the last 24 months
- B. Switch the connection from Extract to Live so Tableau stores no rows locally
- C. Hide all unused fields only after the extract has already been created
- D. Add a context filter on Order Date to the worksheets that use the extract
- E. Enable Aggregate data for visible dimensions with dates rolled up to Day

**Answer:** A, E

**Explanation:** An extract filter limits which rows are written into the extract, and aggregating for visible dimensions with dates rolled to Day stores pre-summarised rows, so both shrink the extract. A context filter acts at query time in a worksheet and does not change what is stored in the extract. Switching to Live removes the extract instead of optimising it, and hiding fields only reduces size if done before the extract is created or followed by a full re-creation, so doing it afterward on its own does not help.

### Question 2 of 10

An analyst connects to an Orders table and a Returns table in Tableau Desktop. Not every order has a return, and the analyst wants Tableau to choose the join type automatically per visualization without duplicating order-level measures. What should the analyst do in the Data Source page?

- A. Create a full outer join between the two tables on Order ID in the physical layer
- B. Add Returns to the logical layer canvas and relate it to Orders on Order ID
- C. Add Returns as a separate data source and blend it as the secondary source linked on Order ID
- D. Open Orders in the physical layer and create a left join to Returns on Order ID

**Answer:** B

**Explanation:** A relationship in the logical layer lets Tableau build the right join per viz at query time and keeps each table at its native level of detail, so Orders measures are not duplicated. A left or full outer join in the physical layer fixes one join type for every sheet and can repeat order rows when there are several returns. Blending also works across sources but is a per-sheet, aggregated technique and is not needed when both tables are in one connection.

### Question 3 of 10

An analyst writes the calculation [Profit] / [Sales] and drags it to a view by Region. The ratio values look much too large. Which calculation returns the correct profit ratio per region?

- A. ATTR([Profit]) / ATTR([Sales])
- B. SUM([Profit]) / SUM([Sales])
- C. SUM([Profit] / [Sales])
- D. [Profit] / SUM([Sales])

**Answer:** B

**Explanation:** Aggregating numerator and denominator separately, SUM(Profit)/SUM(Sales), computes the ratio at the level of the view. A row-level ratio that is then summed adds up thousands of individual ratios, which explains the inflated numbers. ATTR returns a value only when every row shares it, so across many orders it returns an asterisk instead of a ratio, and mixing a row-level field with an aggregate in one expression is not allowed and returns an error.

### Question 4 of 10

A view shows SUM(Sales) by Category and Year. The analyst writes RUNNING_SUM(SUM([Sales])) with Compute Using set to Specific Dimensions, checking both Year and Category with Category listed first. What is the result?

- A. A running total that restarts at the start of each category and only covers that category's years
- B. A running total that adds the categories within each year and then restarts at the start of every new year
- C. A running total that ignores the order of the checked dimensions and sums the table from top to bottom
- D. A running total that moves through every year within one category before continuing to the next category

**Answer:** D

**Explanation:** When both dimensions are addressed, the order of the checked list defines the path, so Category first then Year means the running total goes through all years of one category and continues into the next without restarting. Restarting by category would require Category to be a partition or the restart setting. Adding categories within each year is the reverse order, and the order matters, so it is not simply the visual top-to-bottom order.

### Question 5 of 10 (choose 2)

A dashboard needs to show what share of the total each product contributes while the analyst ranks products from highest to lowest sales. Which two table calculations should be applied? (Choose two.)

- A. Difference on SUM(Sales) computed along Product
- B. Percent of Total on SUM(Sales) computed along Product
- C. Moving Average on SUM(Sales) computed along Product
- D. Percentile on SUM(Sales) computed by Cell only
- E. Rank on SUM(Sales) computed along Product

**Answer:** B, E

**Explanation:** Rank orders products by sales, and Percent of Total divides each product's sales by the sum across the products addressed, which together answer both requirements. A moving average smooths neighbouring values and does not show share or rank. Difference compares each product to the previous one, and a percentile computed by cell has only one value per partition, so it returns the same result for every mark.

### Question 6 of 10

A view shows SUM(Sales) by Customer, and the analyst needs the percentage of customers whose sales exceed 1,000, shown as a single number on a KPI sheet with no dimensions. Which calculation works?

- A. COUNTD(IIF(SUM([Sales]) > 1000, [Customer ID], NULL)) / COUNTD([Customer ID]) with no dimensions
- B. SUM(IIF({ FIXED [Customer ID] : SUM([Sales]) } > 1000, 1, 0)) / COUNTD([Customer ID]) on the KPI sheet
- C. COUNTD(IIF({ FIXED [Customer ID] : SUM([Sales]) } > 1000, [Customer ID], NULL)) / COUNTD([Customer ID])
- D. WINDOW_SUM(IIF(SUM([Sales]) > 1000, 1, 0)) / SIZE() on a KPI sheet that has no dimensions in the view at all

**Answer:** C

**Explanation:** The FIXED LOD returns each customer's total on every row, so counting the distinct IDs that pass the test and dividing by all distinct customers gives the share of customers. Summing a row-level flag counts every order row of a qualifying customer, inflating the numerator. The second COUNTD version mixes an aggregate inside a row-level IIF, which is an error, and the window version has only one mark when no dimension is in the view, so it cannot evaluate customers.

### Question 7 of 10

A designer wants clicking a bar in a Category chart to add or remove that category from a set used to color a line chart, so that multiple categories can be compared over successive clicks. Which action configuration is correct?

- A. A Filter action with Run action on Select and Show all values when clearing selection
- B. A Change Set Values action with Assign values to set and Keep set values when cleared
- C. A Change Set Values action with Add values to set on select and Remove values on clear
- D. A Highlight action on the Category field configured to run on Hover for both sheets

**Answer:** C

**Explanation:** Adding values to the set on each selection accumulates categories, and removing them when cleared lets users reset, so the line chart color comparison builds up over clicks. Assign values replaces membership with the current selection each time instead of accumulating. A filter action removes other categories from the target, and a highlight action only emphasises marks temporarily.

### Question 8 of 10

A dashboard must keep a filter panel and a set of KPI tiles aligned and resizing together when the browser window changes size. What is the BEST way to lay out these objects?

- A. Make every object floating and set exact x and y positions in the Layout pane
- B. Place them inside horizontal and vertical layout containers in the tiled layout
- C. Set the dashboard size to Fixed and add blank objects between the elements
- D. Put each object on its own dashboard and link them with navigation buttons

**Answer:** B

**Explanation:** Layout containers group objects so they resize and stay aligned together, and in a tiled layout they adapt as the dashboard size changes. Floating objects keep fixed positions and can overlap or drift when size changes. Navigation between several dashboards changes the design entirely, and a fixed size with blank spacers avoids resizing rather than handling it.

### Question 9 of 10

A stakeholder wants to compare actual sales against a quota for each sales rep, showing the quota as a marker and actual sales as a bar. Which chart type is designed for this?

- A. Treemap
- B. Area chart
- C. Packed bubble chart
- D. Bullet graph

**Answer:** D

**Explanation:** A bullet graph shows an actual value as a bar against a target reference line and optional performance bands, which is exactly an actual-versus-quota comparison. Packed bubbles and treemaps show part-to-whole size by area and have no built-in target marker. An area chart shows cumulative trends over a continuous axis rather than per-rep targets.

### Question 10 of 10

Sales managers must see only their own region's rows in a published workbook, using a Region Managers table that maps regions to Tableau usernames. Which approach implements row-level security?

- A. A data source filter on a calculation such as [Manager Username] = USERNAME()
- B. A project permission that denies the View capability to other regions' managers
- C. A separate custom view per manager saved with the matching Region selected
- D. A Region filter card on the dashboard with Show Only Relevant Values enabled

**Answer:** A

**Explanation:** Relating the mapping table and filtering the data source on the username field compared with USERNAME() restricts rows per signed-in user, and it cannot be bypassed from the view. A filter card lets users change the selection. Permissions control access to whole content items rather than rows, and custom views store preferences that users can switch away from.

Ready for a timed run? Take the free 15-question SCTDA mini exam: https://devopscodelab.com/app/exams/tableau-data-analyst/mini/

DevOpsCode Lab is not affiliated with Salesforce. SCTDA and related names are trademarks of their owners. We write every question ourselves; these are not real exam questions.
