Tableau Prep practice exam
Fifteen questions across flows, cleaning, reshaping and outputs.
The profile pane shows:
- The flow steps
- Value distributions and nulls per field
- The output path
- Run history
Answer: Value distributions and nulls per field — Data quality at a glance.
While building a flow, Prep works on:
- All data
- A sample
- Metadata only
- The first file
Answer: A sample — Rare values may not appear until you run it.
The join step row count exceeds the left input. This means:
- Normal
- A fan-out from duplicate right-side keys
- Rows were lost
- Wrong output type
Answer: A fan-out from duplicate right-side keys — Aggregate the right branch by the key first.
To see which join values failed to match:
- Run the flow
- Click the unmatched section of the join summary
- Check the output
- Add a clean step
Answer: Click the unmatched section of the join summary — The profile pane then shows them.
Best grouping method for "Smith" and "Smyth":
- Common Characters
- Pronunciation
- Manual
- Spelling
Answer: Pronunciation — Values that sound alike.
The main risk of automatic grouping is:
- Slowness
- Merging genuinely different values
- It cannot be undone
- It only handles numbers
Answer: Merging genuinely different values — Review the proposed groups before applying.
Splitting full names with optional middle names — split from:
- The left
- The right
- The middle
- Automatic only
Answer: The right — The surname is reliably last.
Which is unsupported in Prep?
- FIXED LOD
- Table calculations
- DATEDIFF
- IF/THEN
Answer: Table calculations — They require view context.
A new month column appears each month. Use:
- A manual pivot
- A wildcard pivot
- A union
- An aggregate step
Answer: A wildcard pivot — It matches columns that do not exist yet.
For data Tableau will analyse, pivot:
- Rows to columns
- Columns to rows
- Neither
- Both
Answer: Columns to rows — Long form gives one dimension and one measure.
An aggregate step differs from a FIXED LOD because it:
- Is faster
- Changes the grain, removing detail rows
- Works on text only
- Requires a join
Answer: Changes the grain, removing detail rows — FIXED keeps every row.
The cheapest place to filter is:
- The output
- The input step
- A clean step
- An aggregate step
Answer: The input step — Rows never enter the flow.
Your output doubled on rerun because:
- A fan-out
- The write option was Append
- Sampling
- Wrong join
Answer: The write option was Append — Use create/replace.
Scheduling flows on Server requires:
- Nothing extra
- Prep Conductor or the command line
- Desktop
- A .hyper output
Answer: Prep Conductor or the command line — Conductor is the licensed option.
Incremental refresh does not capture:
- New rows
- Rows updated or deleted after loading
- Large files
- Text fields
Answer: Rows updated or deleted after loading — Hence a periodic full refresh.