Power BI & DAX Quiz

CALCULATE is the only function that changes filter context, e.g. CALCULATE([Sales], Stores[region] = "South").

SUMX iterates row by row, multiplies, then sums; multiplying two totals gives a wrong number, and SUM accepts only a column.

DIVIDE handles division by zero safely and returns BLANK unless you pass a third argument.

Calculated columns use row context and take memory; measures respond to slicers and visuals, so aggregations usually belong in measures.

Each dimension row (e.g. one store) matches many fact rows (many orders), and filters flow from dimension to fact.

CALCULATE([Sales], SAMEPERIODLASTYEAR('Date'[Date])) shifts the current dates back one year; it needs a proper date table.

Inside CALCULATE, ALL acts as a filter remover, so each product's sales are divided by sales across all products.

Power Query records each transformation step as M code; DAX is used afterwards for measures and calculations in the model.

RELATED follows an existing relationship from the many side to the one side; RELATEDTABLE goes the other way and returns a table.

DISTINCTCOUNT counts unique values; COUNTROWS and COUNT count orders, so repeat customers would be counted more than once.

RLS roles filter the data itself for each user; a slicer can simply be changed by the viewer, so it is not security.

Time intelligence shifts and accumulates over a continuous calendar, so a complete, related date table is required.