PowerBI DAX UDF Library

Namespaced, reusable DAX User-Defined Functions (UDFs) for Power BI. Import functions, call them like built-ins, and keep business logic consistent, documented, and versioned.

What is a DAX UDF?

A DAX UDF is a reusable function you define in DAX (preview) and call from measures/expressions. Benefits:

• Consistency
• Maintainability
• Context control

Signature Pattern

DEFINE 
FUNCTION <Namespace.FunctionName> = (
    param1 : <TYPE> [VAL|EXPR],
    ...
) => <FunctionBody>

Types

• NUMERIC
• STRING
• BOOLEAN
• DATETIME
• TABLE
• VARIANT
• INT64
• DECIMAL
• DOUBLE

Modes

• VAL (default): evaluate before call (eager)
• EXPR: evaluate inside function (lazy, context-aware)

Prerequisites

1. Power BI Desktop (latest)
2. Enable DAX UDF:
   • File → Options → Preview features
   • Enable "DAX user-defined functions"
   • Restart Power BI Desktop

Importing Functions

Option A: DAX Query View (Power BI Desktop)

1. Open DAX Query View
2. Paste the DEFINE FUNCTION block(s) from /src/.../*.dax
3. Save to model

Option B:

Usage

Example Function and Measure

DEFINE
/// Financial.AddTax(amount: NUMERIC VAL) -> NUMERIC
/// Returns amount with 10% tax.
/// Model Requirements: none
FUNCTION Financial.AddTax = (
    amount : NUMERIC
) => amount * 1.10

-- In your model
Total Sales w/ Tax = Financial.AddTax ( [Total Sales] )

EXPR vs VAL (context control)

DEFINE
FUNCTION Stats.CountAllDates_VAL = ( t: TABLE VAL ) =>
    COUNTROWS( CALCULATETABLE( t, ALL('Date') ) )

FUNCTION Stats.CountAllDates_EXPR = ( t: TABLE EXPR ) =>
    COUNTROWS( CALCULATETABLE( t, ALL('Date') ) )

Note: With an external 'Date' filter:

• VAL preserves the filter at call time
• EXPR re-evaluates inside the function and can modify context (e.g., ALL('Date'))

Model Requirements

Each .dax file begins with /// docs including Model Requirements:

• Required tables
• Keys
• Relationships
• Cardinality
• Data quality

⚠️ Important: Ensure your model matches these requirements, or outputs may be incorrect.

Conventions

Documentation (mandatory in each file)

/// <Namespace.FunctionName>
/// Summary: <what it does>
/// Params:
///   param (TYPE [VAL|EXPR]): <meaning>
/// Returns: <TYPE>
/// Model Requirements: <tables/columns/relationships>
/// Example:
///   <one-line measure usage>
/// Notes: <edge cases/perf>

Naming

Format: <Namespace>.<VerbNoun> Examples:

• Financial.AddTax
• DateTime.FiscalYearIndex

Performance Best Practices

1. Be explicit with context:

   • KEEPFILTERS
   • REMOVEFILTERS
   • ALL
   • ALLEXCEPT

2. Prefer set-based patterns

   • Avoid repeated heavy calculations
   • Wrap once in a helper UDF

3. Parameter modes:

   • Use EXPR for context-sensitive parameters (tables/measures)
   • Use VAL for scalars

Contributing

Workflow

1. Fork → create feature branch
2. Add function under the correct namespace folder in /src
3. Include complete /// header:
   • Summary
   • Params with type & mode
   • Returns
   • Model Requirements
   • Example
   • Notes
4. Validate in a sample model (e.g., Contoso/AdventureWorks)
5. Open a PR with:
   • Purpose
   • Usage
   • Trade-offs
   • Performance notes (timings/screens if helpful)

PR Checklist

• File name = full function name; correct namespace folder
• /// header complete (incl. Model Requirements)
• Compiles and runs in a basic model
• No breaking changes (or clearly documented)

Folder Additions

When adding a new namespace:

• Include a short README.md inside the folder
• List all functions
• Document typical model prerequisites