2.5.3. Parameters and Dynamic Expressions

💡 First Principle: Parameters and expressions make pipelines flexible and reusable—like variables in programming. Instead of creating separate pipelines for each region (copy-paste nightmare), a single parameterized pipeline handles all regions. The parameter says "which region?" and the pipeline adapts.

Scenario: A pipeline processes sales data for different regions. Instead of creating separate pipelines for each region, use a parameter to specify the region at runtime.

Pipeline Parameters

• Definition: Variables defined at pipeline level
• Usage: Referenced in activities using @pipeline().parameters.ParameterName
• Benefit: Single pipeline serves multiple scenarios

Dynamic Expressions

• Syntax: @{expression} for string interpolation
• Functions: concat(), utcnow(), formatDateTime(), etc.
• Example: Dynamic file path based on date

// Dynamic file path expression
@concat('sales/', formatDateTime(utcnow(), 'yyyy/MM/dd'), '/data.parquet')
// Result: sales/2026/01/02/data.parquet

Visual: Parameterized Pipeline

⚠️ Exam Trap: Creating separate pipelines for each dataset variation is a maintenance nightmare. Use parameters and expressions to create flexible, reusable pipelines. Questions asking for the "most efficient" approach are often testing parameterization knowledge.

Key Trade-Offs:

• Parameterization vs. Simplicity: Heavily parameterized pipelines are flexible but harder to understand and debug
• Dynamic vs. Static Paths: Dynamic paths enable automation but can fail silently if expressions produce unexpected results

Reflection Question: Your pipeline needs to process data from folders named by date (e.g., /2026/01/02/). How would you use expressions to construct the folder path dynamically?