Microsoft PL-300: Prepare the data Practice Questions

DirectQuery is correct because it issues a live query to the source database every time a visual is rendered, ensuring users always see the most current data without importing or caching data. Import is wrong because it loads a snapshot at refresh time and data becomes stale between scheduled refreshes. Dual is wrong because it is a table-level setting used in composite models to allow a table to act as either Import or DirectQuery depending on context — it does not guarantee always-current data for every query. Live Connection to a Power BI semantic model is wrong because that connects to an already-published semantic model in Power BI service, not directly to a SQL Server transactional database.

Dual mode allows a table to behave as an Import table when queries can be satisfied from the cache, and as a DirectQuery table when it is joined to a DirectQuery table — exactly the described requirement. Import alone would break query folding to the DirectQuery source when the tables are joined. DirectQuery on the dimension would always pass queries to the source, losing the performance benefit for slicers. Live Connection is a connection type to an Analysis Services or Power BI shared dataset, not a storage mode for individual tables.

DirectQuery is correct because it sends queries directly to the source database at report render time and stores no data in the Power BI model, satisfying the 'always live' requirement against Azure Synapse Analytics. Import is wrong because it caches a snapshot of the data in the Power BI model and requires scheduled refreshes to stay current. Dual is wrong because it is a per-table setting used in composite models to allow a table to act as either Import or DirectQuery depending on which other tables are being queried—it still caches data. Live connection is wrong because it connects to an existing Analysis Services model or a Power BI shared semantic model, not directly to an operational database like Azure Synapse Analytics.

DirectQuery is correct because it sends queries directly to the source at report render time, ensuring data is always current without importing rows. Import is wrong because scheduled refresh cannot realistically run every minute in the Power BI service (minimum is 30 minutes for most capacities), and it copies data rather than querying live. Dual is wrong because Dual tables still require a DirectQuery partner table; they do not independently guarantee always-current data for a fact table. Live Connection to Azure Analysis Services is wrong because it requires a pre-built AAS model and does not directly connect to Synapse tables.

Column Quality is correct because it shows the percentage of valid, error, and empty values per column — exactly what is needed to identify columns with high error or empty rates. By default, Power Query profiles only the top 1,000 rows; clicking the status bar text 'Column profiling based on top 1000 rows' and switching to 'Column profiling based on entire data set' ensures statistics reflect all 800,000 rows. Column Distribution is incorrect because it shows unique vs. distinct value counts and value distribution, not error/empty percentages. Column Profile shows detailed statistics for a single selected column, not a side-by-side overview for all columns. Increasing the preview row limit in Query Options is not a valid step; the profiling basis is changed via the status bar, not Query Options.