Tableau Data Connections and Preparation 3 — Questions and Answers
Question 1: In Tableau, what is a 'cross-database join' and when would you use it?
- A join between two tables in the same database using different schemas
- A join that combines tables from two different data sources of different types (Correct answer)
- A join performed only on cloud-based databases
- A join that uses a Tableau bridge to connect on-premises data
Correct answer: A join that combines tables from two different data sources of different types
A cross-database join lets you combine tables from entirely different data source types (e.g., SQL Server and Excel) in a single data connection.
Question 2: Which of the following best describes a 'physical table' in Tableau's data model?
- A table that stores only calculated fields
- An actual table or query that contributes data at the join/union layer (Correct answer)
- A virtual table created from a blend
- A summary table generated by Tableau Server
Correct answer: An actual table or query that contributes data at the join/union layer
Physical tables are real tables (or custom SQL queries) that exist at the lower join layer of Tableau's logical data model, where joins and unions are defined.
Question 3: When using custom SQL in Tableau, what is one important performance consideration?
- Custom SQL always runs faster than native table connections
- Tableau cannot create extracts from custom SQL sources
- Custom SQL bypasses Tableau's query optimization and can result in slower performance (Correct answer)
- Custom SQL is only supported for live connections
Correct answer: Custom SQL bypasses Tableau's query optimization and can result in slower performance
Custom SQL is sent as a subquery to the database, which may prevent the database's query optimizer from applying its full optimizations, potentially slowing performance.
Question 4: What does the 'Assume Referential Integrity' option in a Tableau join do?
- Forces Tableau to validate all foreign keys before querying
- Tells Tableau to skip checking for unmatched rows, potentially improving performance (Correct answer)
- Automatically creates primary key constraints in the source database
- Converts inner joins to full outer joins for safety
Correct answer: Tells Tableau to skip checking for unmatched rows, potentially improving performance
'Assume Referential Integrity' instructs Tableau to trust that all join keys match, skipping extra filtering and potentially generating more efficient SQL queries.
Question 5: Which file format does Tableau use when saving a data extract?
- .tde
- .hyper (Correct answer)
- .csv
- .parquet
Correct answer: .hyper
Tableau uses the .hyper file format (introduced in Tableau 10.5) for data extracts, replacing the older .tde format.
Question 6: When you add a second data source to a Tableau workbook and link it via data blending, what field acts as the linking key?
- Any calculated field you define manually
- A field with the same name that is used in the view from the primary source (Correct answer)
- The first date field found in the secondary source
- A field you explicitly drag from the secondary source onto the primary source
Correct answer: A field with the same name that is used in the view from the primary source
Data blending automatically uses fields with matching names between the primary and secondary source as the linking key when those fields appear in the view.
Question 7: A Tableau user is connecting to a large SQL database and wants to limit initial data load for testing purposes. Which option best achieves this?
- Use a live connection with no filters
- Apply an initial SQL statement to truncate the table
- Use extract filters or sampling options during extract creation (Correct answer)
- Switch to a web data connector
Correct answer: Use extract filters or sampling options during extract creation
When creating an extract, Tableau allows you to apply extract filters and set sampling (top N rows or random sample) to limit how much data is loaded.
Tableau Data Connections and Preparation 3
In Tableau, what is a 'cross-database join' and when would you use it?