Salesforce Certified Tableau Desktop Foundations — Practice Set 5

Question 1

A workbook using a live connection is opened while the corresponding database is temporarily unreachable. An identical workbook using an extract of the same data, refreshed the day before, is opened at the same time. What is the key difference in behavior?

Accepted Answer

This is the core live-vs-extract distinction: a live connection depends on ongoing access to the source and will fail or show errors if that source is unreachable, whereas an extract is a self-contained snapshot that works independently of the original source's availability. Tableau does not automatically cache live connections to survive outages, and extracts do not re-validate against the source merely by opening the workbook.

Question 2

Two tables — Orders (grain: one row per order line) and Order Notes (grain: one row per note, with many notes possible per order, and some orders having zero notes) — are combined using an inner join on Order ID. What subtle problem can this introduce compared to using a relationship?

Accepted Answer

This is a classic join pitfall: because a join physically flattens two tables to a single combined grain, an order with three notes produces three duplicated order rows (inflating any SUM on order-level measures), and an inner join drops orders with zero matching notes entirely. A relationship avoids this by preserving each table's native grain and computing aggregations at the appropriate context per worksheet — Tableau does not auto-convert a join to a relationship, joins do not require equal row counts, and there's no requirement to make Order Notes a live connection.

Question 3

A field is renamed from 'cust_id' to 'Customer ID' in the data pane, and separately an alias is applied to the value '1001' so it displays as 'Acme Corp'. What is the key distinction between these two operations?

Accepted Answer

Renaming a field changes the label used to refer to the field itself, wherever it's placed in the workbook, while an alias changes only the display of one specific data value within that field (e.g., '1001' shows as 'Acme Corp') without touching the field's own name or the underlying raw value. Neither operation modifies the raw stored data. Renaming, once applied to a field in the data source, is workbook-wide, not per-worksheet, and aliases are commonly applied to dimension member values, not restricted to measures.

Question 4

A field storing dates as text (e.g., '03/14/2024') is changed to the Date data type, and the values display correctly. Separately, a numeric ZIP code field is assigned the ZIP Code geographic role. What is the key difference between these two actions?

Accepted Answer

Changing a field's data type (e.g., text to Date) fundamentally changes how Tableau parses, validates, and offers functions for that field. Assigning a geographic role is a separate, additive layer that tells Tableau how to interpret an existing field (often still numeric or string) for mapping purposes, without altering its base data type. These are not interchangeable actions, geographic role assignment does not force a 'Geography' data type, and data type changes are commonly needed precisely for text fields being converted to dates or numbers.

Question 5

Which of the following are true distinctions between a live connection and an extract? (Choose two.)

Accepted Answer

The genuine distinctions are that live connections stay current (dependent on network access) while extracts are frozen until refreshed, and extracts are portable/offline-capable while live connections need ongoing source access. It is not universally true that live connections render faster — extracts are often faster since Tableau's optimized in-memory engine avoids repeated source queries. Extracts absolutely can, and typically do, contain calculated fields.

Question 6

A data source contains a Transaction Amount field stored as a string type. The author needs to sum this field but the SUM aggregation is unavailable in the Marks card. What is the most likely underlying issue?

Accepted Answer

Numeric aggregations like SUM require the field to be recognized as a number data type; a field still typed as a string will not offer SUM until its data type is converted. Geographic roles and hierarchies are unrelated to enabling numeric aggregation, and SUM is available for fields converted from any originating format once their data type is numeric — it's not permanently blocked by the source format.

Question 7

An author creates a Set for 'Customers with Sales > $10,000' and, separately, a Group combining 'CA', 'OR', and 'WA' into 'West Coast'. Later, the underlying data changes so that a previously excluded customer now has sales of $12,000, and a new state value 'NV' is added to the data intended to also belong to the West Coast group. What is the key behavioral difference between the set and the group in this situation?

Accepted Answer

A condition-based set recalculates its membership dynamically whenever the underlying data changes and a record newly satisfies (or stops satisfying) the condition — so the customer now exceeding $10,000 is automatically included. A group, by contrast, is a static, manually curated list of specific existing members; a brand-new value like 'NV' will not automatically join the 'West Coast' group and must be added manually by the author. This distinction is reversed from options C and D, and both can in fact be updated (manually for groups, automatically for condition-based sets).

Question 8

An author applies a filter on Region directly to the Filters shelf on one worksheet, and separately uses a parameter with a calculated field to simulate filtering logic on another worksheet. What is a key functional difference between these two approaches?

Accepted Answer

A true filter on the Filters shelf excludes non-matching data from the underlying query, reducing what's actually retrieved and displayed. A parameter referenced in a calculated field, by contrast, typically drives a computed value or a TRUE/FALSE flag used for coloring, highlighting, or conditional formatting — the excluded-looking rows are often still present in the data, just displayed differently, unless that calculated field is also explicitly placed on the Filters shelf. Parameters can be used with calculations referencing measures, and filters are available directly on worksheets, not just dashboards.

Question 9

A worksheet shows Order Date as a discrete 'Month' on Columns. A second version of the same worksheet shows Order Date as a continuous 'Month' on Columns. Both worksheets are filtered to show only 2023 and 2024 data. What is the key visual gotcha to be aware of?

Accepted Answer

This is a well-known gotcha: a discrete date part (like 'Month') groups all matching periods together regardless of year, collapsing January 2023 and January 2024 into one 'January' header — which can misleadingly combine two years of data into a single bar. A continuous date value, by contrast, plots each month/year combination as its own separate point along an unbroken timeline. The discrete version can still be filtered by year (the filter just doesn't stop the collapsing behavior for the month header), and it's the discrete treatment — not continuous — that merges same-named periods.

Question 10

An author applies a Top 10 Customers by Sales filter directly on the Filters shelf, and separately builds a Top 10 Customers by Sales Set used only for coloring (in-set vs. out-of-set) without placing it on the Filters shelf. If a Region filter is then also applied to the view, what is a key gotcha to watch for?

Accepted Answer

This highlights a genuine gotcha in Tableau: filters on the Filters shelf are affected by context and filter order relative to other filters (e.g., a Top N filter can be computed before or after a Region filter depending on context filter settings), while a Set's computed membership (such as Top N by Sales) is calculated independently based on how the set itself was defined, and won't automatically respect other filters unless the set condition or a context filter is explicitly configured to do so. Sets and filters do not automatically share identical evaluation scope, applying one filter doesn't disable another, and sets absolutely can be defined using Top N conditions.

Question 11

A hierarchy is built from Category > Sub-Category, and a separate hierarchy exists for Region > State > City. An author wants a single worksheet where clicking a bar drills from Category into Sub-Category, and independently drills from Region into State into City, both governed by their own expand/collapse controls. Is this possible, and why?

Accepted Answer

A worksheet can host multiple independent hierarchies simultaneously (for example, one on Columns and one on Rows, or nested at different points in the view), each retaining its own expand/collapse drill behavior. There's no one-hierarchy-per-worksheet limitation, hierarchies are built from dimensions (like Region, State, City), not measures, and there's no requirement to merge separate hierarchies into one to use them together.

Question 12

A relative date filter is set to 'Last 4 Quarters' and a second, separate filter uses a fixed range covering the same historical four quarters as of today. Six months from now, what is the key behavioral difference?

Accepted Answer

This is the defining behavioral gap between the two filter types: a relative date filter recalculates its window relative to whenever the workbook is viewed, so 'Last 4 Quarters' will always mean the four quarters ending now. A fixed range filter locks in specific calendar dates and will not move forward on its own, becoming stale over time unless someone manually edits it. Relative date filters do support quarter-based (and many other) granularities, not just days or months.

Question 13

Which of the following correctly distinguish a bin from a group in Tableau? (Choose two.)

Accepted Answer

Bins are specifically for segmenting continuous numeric ranges into equal-width buckets (e.g., Age into 0-9, 10-19), while groups combine specific existing categorical dimension members (e.g., three states into one region label) without requiring numeric interval logic. They are not interchangeable — a bin wouldn't make sense for combining arbitrary text categories, and a group wouldn't create equal-width numeric buckets. Groups are very commonly based on text dimensions (like state names), not restricted to numeric measures.

Question 14

A calculated field uses a parameter to let users toggle between showing SUM(Sales) or SUM(Profit) in the same chart. Separately, a different calculated field hardcodes a fixed threshold value (e.g., 5000) with no parameter. What is the key tradeoff of the hardcoded version?

Accepted Answer

This is a key design tradeoff tested at the foundations level: a hardcoded value baked into a calculation can only be changed by editing that calculation (and republishing, if applicable), whereas a parameter lets end users adjust the value dynamically through an interactive control, with no need to edit the underlying calculation. A hardcoded field does not retroactively become parameter-driven just because a parameter is added elsewhere, and parameters are commonly used for both measure-swapping and adjustable numeric thresholds — the two use cases are not mutually exclusive.

Question 15

A crosstab shows Show Totals enabled for Row Grand Totals only, not Column Grand Totals. A second, similar crosstab has both enabled. What is the key distinction in what is displayed?

Accepted Answer

Row Grand Totals and Column Grand Totals are independently toggleable options within Show Totals, so enabling only one produces a summarizing total in just that direction (e.g., at the end of each row), while enabling both adds totals in both directions of the crosstab. They do not automatically show identical values (they summarize different dimensions of the table), enabling one does not remove existing subtotals, and Show Totals is very much applicable to crosstabs, not limited to bar charts.

Question 16

A dashboard uses a Highlight Action triggered by hover on Worksheet A to emphasize related marks on Worksheet B. A second dashboard uses a Filter Action triggered by click on Worksheet A to filter Worksheet B instead. If a user hovers (without clicking) over a mark on Worksheet A in each dashboard, what is the key behavioral difference?

Accepted Answer

Each dashboard action's trigger (hover, select/click, or menu) is configured independently, so a highlight action set to trigger on hover reacts immediately to hovering, while a filter action set to trigger on click requires an actual click and does nothing on mere hover. Highlight and filter actions do not share a single fixed trigger by default, filter actions can be configured to trigger on hover or click (author's choice), and highlight actions only visually de-emphasize non-matching marks — they do not remove them from the view.

Question 17

An author builds a story with five story points capturing different filter states of the same dashboard. Later, the author edits the dashboard's underlying worksheet to add a new field to the Marks card. What happens to the previously captured story points?

Accepted Answer

Story points capture a particular state (such as filter selections and view configuration) of the referenced worksheets/dashboards, not a frozen static image — so when the underlying dashboard is edited (like adding a new field to Marks), that change is reflected the next time each story point is viewed, while the captured navigation/filter state itself remains as originally set. Story points are not deleted by unrelated dashboard edits, and there's no restriction requiring all edits to occur before creating story points.

Question 18

A dashboard has both a default (desktop) layout and a phone-specific layout configured. An author edits a worksheet by adding a new filter card to the default layout only, not touching the phone layout. What is a likely gotcha here?

Accepted Answer

Device-specific layouts (like phone) are maintained somewhat independently once customized — an object added only to the default layout does not automatically propagate to a phone layout that has already been separately configured, so the author must add it there as well if it should also appear on phone. Editing the default layout does not delete the phone layout, and device-specific layouts can absolutely contain filter cards like any other dashboard object.

Question 19

An author exports a dashboard to PowerPoint, and separately saves the same workbook as a packaged workbook (.twbx) to share with a colleague. What is the key functional difference between what each recipient can do?

Accepted Answer

Exporting to PowerPoint captures the current visual state as static slide content, suitable for presentations but without click-to-filter or hover interactivity, whereas a .twbx file bundles the full workbook (and any extract data) so the recipient can interact with filters, actions, and tooltips when opened in Tableau Desktop or the free Tableau Reader. A .twbx does not need to be exported to PowerPoint to be opened, and PowerPoint exports do not embed the full underlying data as an editable table — they are visual snapshots.

Question 20

A user views the underlying data behind a mark that represents an aggregated SUM(Sales) value across 50 individual transactions, then exports that view to .csv. What does the exported file contain?

Accepted Answer

Viewing (and then exporting) the underlying data for a mark surfaces the individual records behind that aggregation — in this case, the 50 disaggregated transaction rows, not merely the one summed total value. It does not export calculated field formulas as executable logic, and this differs entirely from exporting a static image, which would be a visual snapshot, not row-level data.

Question 21

Which of the following correctly distinguish a Filter Action from a Highlight Action when clicking a mark on a source worksheet? (Choose two.)

Accepted Answer

A filter action truly restricts what data is retrieved/shown on the target worksheet to matching records, while a highlight action keeps all the original data visible but visually emphasizes the matching subset (commonly by dimming the rest) — a much less destructive interaction. These two produce clearly different visual results, and a highlight action never deletes data from the data source; it only affects the visual emphasis in the current view.

Question 22

A field named 'Transaction Count' is a plain integer field where each row already represents one transaction (i.e., the value is always 1). An author uses SUM(Transaction Count) in one worksheet and COUNTD(Transaction ID) in another, both intended to answer 'how many transactions occurred.' Under what circumstance would these two produce different results?

Accepted Answer

This tests a subtle measure/aggregation gotcha related to the earlier join-vs-relationship duplication risk: if a join duplicates rows (each transaction now appearing multiple times), SUM(Transaction Count) sums the duplicated 1s and over-counts, whereas COUNTD(Transaction ID) counts only distinct transaction IDs and remains accurate despite duplicated rows. The two aggregations are not always equivalent, COUNTD works perfectly well on text/ID fields (a very common use case), and SUM on a numeric 'Transaction Count' field is a perfectly valid aggregation.

Question 23

A field 'Rating' contains values 1 through 5, representing a customer satisfaction rating. An author debates whether to treat it as a discrete or continuous field for a bar chart segmenting average sales by rating. What is the practical consideration?

Accepted Answer

This reflects the genuine, practical discrete-vs-continuous decision tested at the foundations level: since Rating has a small number of distinct, meaningful categories (1-5), treating it as discrete yields separate bar headers per rating — appropriate for category-style comparison — while continuous treatment would place it on a numeric axis better suited to trend or distribution-style visuals rather than discrete category bars. The two treatments are visually distinct, not identical. Numeric fields like Rating can be used as either a measure or (commonly, when discrete) as a dimension depending on intended use, and discrete fields — not exclusively continuous ones — are what typically segment a bar chart into separate category bars.

Question 24

Which statement correctly describes the default aggregation behavior when a numeric measure like Sales is first dragged into a new worksheet?

Accepted Answer

Tableau applies a default aggregation, most commonly SUM (though this can be changed via the field's default properties), automatically when a measure is placed into the view — the author is not required to manually pick an aggregation before the field functions at all. Dimensions are typically not aggregated (they segment data instead), and the default aggregation is not fixed to AVG; it depends on the field's configured default properties, which are often SUM unless changed.

Question 25

A worksheet shows AVG(Order Value) with Region on Rows. The author then adds Customer Segment to Rows as well, nested under Region. The average shown for a given Region/Segment combination is different from the average shown for that Region alone (with no segment breakdown). Why is this expected behavior rather than an error?

Accepted Answer

This is the same underlying level-of-detail principle tested throughout the exam, extended to nested dimensions: each additional dimension added to the view further refines the granularity at which an aggregated measure is computed, so the Region-and-Segment average is legitimately expected to differ from the broader Region-only average — this is normal aggregation behavior, not a data quality problem. The aggregation function itself (AVG) does not automatically change to SUM, and Customer Segment remains a dimension; it is not converted into a measure by being nested in the view.

Analytics-101 — Salesforce Certified Tableau Desktop Foundations