In a Fabric semantic model using Direct Lake mode, which tool can identify frequently used columns loaded into memory?

In a Fabric semantic model using Direct Lake mode, which tool can identify frequently used columns loaded into memory?

• A. Vertipaq Analyzer
• B. Query the $System.DISCOVER_STORAGE_TABLE_COLUMN_SEGMENTS dynamic management view (DMV)
• C. Data Size Monitor
• D. Power BI Performance Analyzer

Answer: (A)

In Direct Lake mode, you want to see which columns the engine actually keeps in memory during query processing so you can optimize what’s loaded. VertiPaq Analyzer is built to inspect the model’s in‑memory structure and show memory usage at the column level, including which columns are loaded, how much memory they consume, and how compression and data types affect that usage. This makes it the best tool for identifying frequently used columns that are loaded into memory, enabling targeted optimizations for performance and resource use. Other options exist for diagnostics, but they’re less suited for quickly pinpointing column‑level memory loading. The DMV approach exposes memory details at a low level, which is powerful but more manual and technical. Data Size Monitor tracks overall data size rather than per‑column memory loading, and Power BI Performance Analyzer focuses on report and client performance rather than the model’s internal memory footprint.

In Direct Lake mode, you want to see which columns the engine actually keeps in memory during query processing so you can optimize what’s loaded. VertiPaq Analyzer is built to inspect the model’s in‑memory structure and show memory usage at the column level, including which columns are loaded, how much memory they consume, and how compression and data types affect that usage. This makes it the best tool for identifying frequently used columns that are loaded into memory, enabling targeted optimizations for performance and resource use.

Other options exist for diagnostics, but they’re less suited for quickly pinpointing column‑level memory loading. The DMV approach exposes memory details at a low level, which is powerful but more manual and technical. Data Size Monitor tracks overall data size rather than per‑column memory loading, and Power BI Performance Analyzer focuses on report and client performance rather than the model’s internal memory footprint.