PL-300 Question #170: Real Exam Question with Answer & Explanation

Question

You have an imported dataset that contains the following tables: Date Product Customer SalesFact You need to reduce the size of the dataset before you publish it. The solution must minimize the number of tables in the dataset. What should you do?

Options

• AMark the Date table as the date table for the model.
• BChange the storage mode of all the tables to DirectQuery.
• CReplace DAX calculated columns with Power Query equivalents.
• DSummarize data at the month level in the SalesFact query.

Explanation

Replacing DAX calculated columns with Power Query equivalents reduces dataset size because DAX calculated columns are computed after data loads and stored separately in the VertiPaq in-memory engine, increasing the model's memory footprint. Power Query columns are computed during the ETL stage and compressed together with the base table data - making them more storage-efficient and keeping everything within existing tables, satisfying the "minimize table count" constraint.

Why the distractors are wrong:

• A - Marking the Date table as the date table is a semantic/relationship setting; it has no effect on dataset size or table count.
• B - Switching to DirectQuery means data isn't imported, but it doesn't reduce the number of tables and introduces query performance and connectivity tradeoffs - it's a storage mode change, not a size optimization.
• D - Summarizing SalesFact at the month level would reduce row count, but it doesn't minimize table count (you may even need an additional aggregated table), and you risk losing granularity needed for reporting.

Memory tip: Think of it this way - DAX calculated columns live inside the model (memory cost), while Power Query columns live inside the query (ETL cost). Moving work earlier in the pipeline (to Power Query) keeps the loaded model lean and avoids bloating the VertiPaq store.

No community discussion yet for this question.