Reduce Power BI Data Model Size Without Losing Information
You import a Power BI dataset that contains the following tables: ✑ Date ✑ Product ✑ Product Inventory The Product Inventory table contains 25 million rows. A sample of the data is shown in the following table. [sampleXML/xmlfile-119_4.png] The Product Inventory table relates to the Date table by using the DateKey column. The Product Inventory table relates to the Product table by using the ProductKey column. You need to reduce the size of the data model without losing information. What should you do?
Community Insight
The core concept tested is data model optimization through column reduction, with the common trap being confusion between summarization settings and actual data storage volume.
This question addresses techniques to optimize a large Power BI data model by identifying and removing redundant columns. It establishes that removing duplicate or unnecessary columns is the most effective way to reduce storage size while preserving all unique data.
Many learners incorrectly choose Option A (Change Summarization for DateKey to Don't Summarize) because they misunderstand that changing aggregation behavior does not reduce the physical storage size of the underlying rows.
Community Discussion (88 comments)
- I said: D
Comments & Corrections
No comments yet — spotted an error or have a note? Share it below.
Expert Analysis
Why the Answer Is Correct
Removing the MovementDate column (Option D) is the correct action because, as noted in community comments, it appears to be a duplicate of DateKey. In Power BI, each column consumes memory based on its cardinality and data type. If two columns contain identical information, one can be safely removed without losing any analytical capability or data integrity. This directly reduces the size of the dataset.Why the Other Options Are Wrong
Option A is incorrect because changing the summarization property affects how measures aggregate data in visuals but does not change the number of rows stored in the column. Options B and C are detrimental: removing relationships breaks data analysis capabilities, and changing UnitCost to Integer would likely result in loss of precision (information loss), violating the 'without losing information' constraint.Community Comment Notes
Community consensus strongly supports Option D. As user ElijahSu noted, "D is right." User semauni clarified the reasoning: "DateKey and MovementDate contain the same data:)". User Snow_28 added that "changing the summarize option... will not affect the size of the model," confirming why Option A is a distractor.Exam Strategy
Always look for redundancy in large tables before considering complex optimizations. If two columns seem to hold the same information, removing one is the fastest way to shrink the model size without affecting functionality.
Frequently Asked Questions
Why doesn't changing summarization reduce model size?
Summarization only changes how data aggregates in visuals; it does not remove rows from the table or reduce the storage footprint.
Is it safe to remove a column if it duplicates another?
Yes, if the columns contain identical values, removing one preserves all unique data and significantly reduces memory usage.
Related Analysis
Practice All PL-300 Questions
Access 116 questions with complete answers and detailed explanations.
View Full PL-300 Practice Test →