DP-600 - Prepare Data (46% of the exam) - Section 2.7

Select, filter, and aggregate data using the Visual Query Editor and SQL against the SQL analytics endpoint.

Write T-SQL SELECT statements using WHERE, JOIN, and GROUP BY aggregation against the SQL analytics endpoint to retrieve and summarise data. Use the Visual Query Editor as a no-code alternative and recognise when each tool is appropriate based on user skill and query complexity.

Visual Query EditorT-SQLSQL analytics endpointGROUP BY aggregationresult set filtering

Practice question for this objective

Free samplePrepare Datamedium

An analyst who prefers a no-code experience opens the "Visual Query Editor" on a "SQL analytics endpoint" to build a summary that counts orders per status. They drag the orders table onto the canvas and now need to produce one row per status with a row count. Which step in the Visual Query Editor produces this aggregation, and what does the editor generate underneath?

  • AApply the Filter rows transformation on status, which collapses duplicate status values into one row each and emits a SELECT DISTINCT statement under the canvas.
  • BApply the Group by transformation, choosing status as the grouping column and a count aggregation, and the editor generates the equivalent T-SQL with a GROUP BY clause. Correct
  • CApply the Sort transformation on status so identical values sit together, and the editor emits an ORDER BY statement that the report then counts visually per block.
  • DSwitch to the Merge queries transformation to join the table to itself on status, and the editor emits a self-join that totals the matching rows per status.
Aggregate in the Visual Query Editor with the Group by step, which generates the equivalent GROUP BY T-SQL for the SQL analytics endpoint. The Visual Query Editor is a diagram-based surface where each transformation maps to a SQL operation; the Group by step corresponds to a GROUP BY with an aggregate, so choosing status as the key and count as the aggregate yields one counted row per status and produces the matching T-SQL beneath the canvas.

Why A is wrong: Filtering rows removes records that fail a condition rather than aggregating, so it neither produces per-status counts nor generates the grouping logic the task needs.

Why B is correct: The Group by step is the visual equivalent of GROUP BY, and the editor renders the diagram as T-SQL behind the canvas, so it both aggregates correctly and produces reviewable SQL.

Why C is wrong: Sorting only orders rows and produces ORDER BY, which does not compute counts per status, so the aggregation still has to be done somewhere else.

Why D is wrong: Merge queries performs a join between tables rather than an aggregation, so a self-join overcomplicates the task and does not cleanly yield one counted row per status.

See more DP-600 practice questions, answers explained.

Exam traps in Prepare Data

Answers that look right on this material and are not. Each one is a distractor from a different question in the DP-600 bank for this domain.

  • A new "Warehouse" built beside the "Lakehouse", which the analyst queries with read-only T-SQL and views while leaving the original Delta tables untouched.

    Why it is wrong: A "Warehouse" is a separate read-write store that would require copying or referencing data into a new item, which adds an unnecessary store when a read endpoint already exists.

  • Use the Append queries step on the two tables, and the editor generates a UNION ALL that stacks the customers table beneath the orders table to add the name column.

    Why it is wrong: Append stacks rows from two tables of the same shape and emits UNION ALL, so it adds rows rather than matching on a key to attach the customer name to each order.

  • The endpoint blocks UPDATE only on tables larger than the capacity row limit, so the engineer should split the table into smaller partitions first and then rerun the same UPDATE statement.

    Why it is wrong: There is no row-count threshold that turns write access on; the endpoint forbids data modification regardless of table size, so partitioning would not unlock an UPDATE.

Examworthy is not affiliated with or endorsed by Microsoft. Original, blueprint-aligned practice material only.