Optimize Model Performance PL-300 Practice Questions (Page 4)

Questions 16–20

1. 16application · medium

   A retail company's Power BI model imports a product table with 50 columns. The report only uses 'ProductID', 'ProductName', 'Category', and 'Price'. The table also contains a 'ProductDescription' column with long text that is not used. The model refresh is slow. What should you do to improve refresh performance while keeping the report functional?

   • ARemove all columns except the four used in the report, including 'ProductDescription'.
   • BKeep all columns but change 'ProductDescription' to a numeric data type.
   • CCreate a calculated column that extracts the first 10 characters of 'ProductDescription' and use that instead.
   • DDisable the auto date/time feature and keep all columns.
   Reveal Solution Discussion Select an answer first
2. 17application · medium

   A measure 'Total Profit' is defined as SUMX(Sales, Sales[Revenue] - Sales[Cost]). The visual using this measure is slow. You analyze the DAX query and see that the SUMX iterates over many rows. What should you do to improve performance?

   • AUse a visual filter to limit the rows that SUMX iterates over.
   • BReplace the measure with a calculated column for profit and then sum that column.
   • CAdd more columns to the Sales table to give SUMX more data to work with.
   • DChange the measure to use SUM(Sales[Revenue]) - SUM(Sales[Cost]).
   Reveal Solution Discussion Select an answer first
4. 18application · medium

   A manufacturing company imports a sensor table that records temperature every second for each machine. The report needs to show average temperature per hour per machine. The table also includes a 'SensorID' column that is not used in any report. The model is too large and refresh takes too long. What should you do to reduce the model size while still supporting the report?

   • AKeep the raw sensor data but remove the 'SensorID' column and create a measure for average temperature.
   • BUse a row-level security (RLS) rule to limit the sensor data to the current hour for all users.
   • CImport the raw sensor data and use a calculated table to create hourly averages.
   • DAggregate the sensor data to hourly averages per machine before importing, and remove the 'SensorID' column.
   Reveal Solution Discussion Select an answer first
5. 19expert · hard

   A measure 'Total Sales' is defined as CALCULATE(SUM(Sales[Amount]), Sales[Region] = "West"). The measure is slow. You use DAX Studio to analyze the query and see that the storage engine is scanning the entire Sales table. What is the most likely cause?

   • AThe filter on Sales[Region] is not being pushed down to the storage engine because the measure uses CALCULATE.
   • BThe Sales[Amount] column is not properly compressed, causing a full scan.
   • CThe Sales[Region] column has a high cardinality, making the filter inefficient.
   • DThe Sales table has too many rows, so the scan takes longer.
   Reveal Solution Discussion Select an answer first
6. 20expert · hard

   A DAX measure 'Sales YTD' uses CALCULATE(SUM(Sales[Amount]), DATESYTD('Date'[Date])). The measure is slow when used in a visual with a large date range. You analyze the DAX query and see that the storage engine is scanning the entire Sales table. What should you do to improve performance?

   • AIncrease the number of rows in the Sales table to improve the scan efficiency.
   • BReplace DATESYTD with a custom date filter that uses a calculated column for the year-to-date flag.
   • CRemove the date table and use the date column from the Sales table directly.
   • DUse a visual filter to limit the date range to one year.
   Reveal Solution Discussion Select an answer first