Five DAX Patterns Every Power BI Analyst Should Know

Pattern 1 — Measures over calculated columns

As a rule: prefer measures. A calculated column is computed once and stored in every row of your model, whether or not a report ever uses it. A measure is computed on the fly, in the filter context of whatever visual is displaying it — which is also what makes it correct when a user slices by a different dimension.

Total Revenue = SUM ( Sales[Amount] )

That single measure works whether the visual is sliced by country, product, or nothing at all. A calculated column trying to do the same job would need to guess the filter context in advance, which it can't.

Pattern 2 — Year-to-date

Revenue YTD =
CALCULATE (
    [Total Revenue],
    DATESYTD ( 'Date'[Date] )
)

CALCULATE is DAX's context-modifier function — it takes an existing measure and evaluates it under a different filter than the one the visual would normally apply. DATESYTD generates that date range automatically, provided you have a proper marked date table (Power BI's automatic date hierarchies don't count — a real date table is worth setting up early).

Pattern 3 — Year-over-year growth

Revenue YoY % =
VAR Current = [Total Revenue]
VAR Prior = CALCULATE ( [Total Revenue], SAMEPERIODLASTYEAR ( 'Date'[Date] ) )
RETURN DIVIDE ( Current - Prior, Prior )

Two things worth copying here regardless of the specific calculation: naming intermediate values with VAR instead of repeating [Total Revenue] three times, and using DIVIDE() instead of the / operator — DIVIDE returns BLANK() on a divide-by-zero instead of an error that breaks the whole visual.

Pattern 4 — Running totals

Running Revenue =
CALCULATE (
    [Total Revenue],
    FILTER (
        ALL ( 'Date'[Date] ),
        'Date'[Date] <= MAX ( 'Date'[Date] )
    )
)

ALL('Date'[Date]) removes the existing date filter so you can rebuild it yourself; the FILTER then keeps every date up to (and including) the current row's date. This is the DAX equivalent of the SQL rows between unbounded preceding and current row window frame — same concept, different syntax.

Pattern 5 — Dynamic ranking with RANKX

Product Rank =
RANKX (
    ALL ( Sales[Product] ),
    [Total Revenue],
    ,
    DESC
)

A table visual sorted by a revenue column looks ranked, until someone filters by region and the sort silently stops reflecting a true rank within that filter. RANKX recalculates against the current filter context every time, so the rank stays correct no matter how the report is sliced.

Summary

A clean, marked date table is the foundation every time intelligence pattern above depends on — it's worth building before writing a single measure.

Where this shows up in practice

The Retail Sales Performance Dashboard project uses RANKX for product ranking directly; the E-commerce Order & Revenue Analytics project uses the same CALCULATE-based pattern for its run-rate revenue forecast.