Spark Declarative Pipelines

Lakeflow Spark Declarative Pipelines (SDP)

Quick Reference

Official Documentation

• Lakeflow Spark Declarative Pipelines Overview - Main documentation hub
• SQL Language Reference - SQL syntax for streaming tables and materialized views
• Python Language Reference - pyspark.pipelines API
• Loading Data - Auto Loader, Kafka, Kinesis ingestion
• Change Data Capture (CDC) - AUTO CDC, SCD Type 1/2
• Developing Pipelines - File structure, testing, validation
• Liquid Clustering - Modern data layout optimization

Quick Start: Initialize New Pipeline Project

RECOMMENDED: Use databricks pipelines init to create production-ready Asset Bundle projects with multi-environment support.

When to Use Bundle Initialization

Use bundle initialization for New pipeline projects for a professional structure from the start

Use manual workflow for:

• Quick prototyping without multi-environment needs
• Existing manual projects you want to continue
• Learning/experimentation

Step 1: Initialize Project

I will automatically run this command when you request a new pipeline:

databricks pipelines init

Interactive Prompts:

• Project name: e.g., customer_orders_pipeline
• Initial catalog: Unity Catalog name (e.g., main, prod_catalog)
• Personal schema per user?: yes for dev (each user gets their own schema), no for prod
• Language: SQL or Python (auto-detected from your request - see language detection below)

Generated Structure:

my_pipeline/
├── databricks.yml              # Multi-environment config (dev/prod)
├── resources/
│   └── *_etl.pipeline.yml      # Pipeline resource definition
└── src/
    └── *_etl/
        ├── explorations/       # Exploratory code in .ipynb
        └── transformations/    # Your .sql or .py files here

Step 2: Customize Transformations

Replace the example code created by the init process with custom transformation files in src/transformations/ based on provided requirements, using best practice guidance from this skill.

Step 3: Deploy and Run

# Deploy to workspace (dev by default)
databricks bundle deploy

# Run pipeline
databricks bundle run my_pipeline_etl

# Deploy to production
databricks bundle deploy --target prod

I can run these commands for you using the Bash tool.

For medallion architecture (bronze/silver/gold), two approaches work:

• Flat with naming (template default): bronze_*.sql, silver_*.sql, gold_*.sql
• Subdirectories: bronze/orders.sql, silver/cleaned.sql, gold/summary.sql

Both work with the transformations/** glob pattern. Choose based on preference.

See 8-project-initialization.md for complete details on bundle initialization, migration, and troubleshooting.

Alternative: Manual Workflow (Advanced)

For rapid prototyping, experimentation, or when you prefer direct control without Asset Bundles, use the manual workflow with MCP tools.

Use MCP tools to create, run, and iterate on serverless SDP pipelines. The primary tool is create_or_update_pipeline which handles the entire lifecycle.

IMPORTANT: Always create serverless pipelines (default). Only use classic clusters if user explicitly requires R language, Spark RDD APIs, or JAR libraries.

Step 1: Write Pipeline Files Locally

Create .sql or .py files in a local folder:

my_pipeline/
├── bronze/
│   ├── ingest_orders.sql       # SQL (default for most cases)
│   └── ingest_events.py        # Python (for complex logic)
├── silver/
│   └── clean_orders.sql
└── gold/
    └── daily_summary.sql

SQL Example (bronze/ingest_orders.sql):

CREATE OR REFRESH STREAMING TABLE bronze_orders
CLUSTER BY (order_date)
AS
SELECT
  *,
  current_timestamp() AS _ingested_at,
  _metadata.file_path AS _source_file
FROM read_files(
  '/Volumes/catalog/schema/raw/orders/',
  format => 'json',
  schemaHints => 'order_id STRING, customer_id STRING, amount DECIMAL(10,2), order_date DATE'
);

Python Example (bronze/ingest_events.py):

from pyspark import pipelines as dp
from pyspark.sql.functions import col, current_timestamp

@dp.table(name="bronze_events", cluster_by=["event_date"])
def bronze_events():
    return (
        spark.readStream.format("cloudFiles")
        .option("cloudFiles.format", "json")
        .load("/Volumes/catalog/schema/raw/events/")
        .withColumn("_ingested_at", current_timestamp())
        .withColumn("_source_file", col("_metadata.file_path"))
    )

Language Selection:

• Auto-detection: I analyze your request for keywords:
  • SQL indicators: "SQL", "sql files", "simple transformations", "aggregations", "materialized view", "CREATE OR REFRESH"
  • Python indicators: "Python", ".py files", "UDF", "complex logic", "ML inference", "external API", "@dp.table", "pandas"
• Prompt for clarification when language intent is unclear or mixed
• Use SQL for: Transformations, aggregations, filtering, joins (most cases)
• Generate ONE language per request unless you explicitly ask for mixed pipeline

See 8-project-initialization.md for detailed language detection logic.

Step 2: Upload to Databricks Workspace

# MCP Tool: upload_folder
upload_folder(
    local_folder="/path/to/my_pipeline",
    workspace_folder="/Workspace/Users/user@example.com/my_pipeline"
)

Step 3: Create/Update and Run Pipeline

Use create_or_update_pipeline - the main entry point. It:

1. Searches for an existing pipeline with the same name (or uses id from extra_settings)
2. Creates a new pipeline or updates the existing one
3. Optionally starts a pipeline run
4. Optionally waits for completion and returns detailed results

# MCP Tool: create_or_update_pipeline
result = create_or_update_pipeline(
    name="my_orders_pipeline",
    root_path="/Workspace/Users/user@example.com/my_pipeline",
    catalog="my_catalog",
    schema="my_schema",
    workspace_file_paths=[
        "/Workspace/Users/user@example.com/my_pipeline/bronze/ingest_orders.sql",
        "/Workspace/Users/user@example.com/my_pipeline/silver/clean_orders.sql",
        "/Workspace/Users/user@example.com/my_pipeline/gold/daily_summary.sql"
    ],
    start_run=True,           # Start immediately
    wait_for_completion=True, # Wait and return final status
    full_refresh=True,        # Full refresh all tables
    timeout=1800              # 30 minute timeout
)

Result contains actionable information:

{
    "success": True,                    # Did the operation succeed?
    "pipeline_id": "abc-123",           # Pipeline ID for follow-up operations
    "pipeline_name": "my_orders_pipeline",
    "created": True,                    # True if new, False if updated
    "state": "COMPLETED",               # COMPLETED, FAILED, TIMEOUT, etc.
    "catalog": "my_catalog",            # Target catalog
    "schema": "my_schema",              # Target schema
    "duration_seconds": 45.2,           # Time taken
    "message": "Pipeline created and completed successfully in 45.2s. Tables written to my_catalog.my_schema",
    "error_message": None,              # Error summary if failed
    "errors": []                        # Detailed error list if failed
}

Step 4: Handle Results

On Success:

if result["success"]:
    # Verify output tables
    stats = get_table_details(
        catalog="my_catalog",
        schema="my_schema",
        table_names=["bronze_orders", "silver_orders", "gold_daily_summary"]
    )

On Failure:

if not result["success"]:
    # Message includes suggested next steps
    print(result["message"])
    # "Pipeline created but run failed. State: FAILED. Error: Column 'amount' not found.
    #  Use get_pipeline_events(pipeline_id='abc-123') for full details."

    # Get detailed errors
    events = get_pipeline_events(pipeline_id=result["pipeline_id"], max_results=50)

Step 5: Iterate Until Working

1. Review errors from result or get_pipeline_events
2. Fix issues in local files
3. Re-upload with upload_folder
4. Run create_or_update_pipeline again (it will update, not recreate)
5. Repeat until result["success"] == True

Quick Reference: MCP Tools

Primary Tool

Pipeline Management

Supporting Tools

Reference Documentation (Local)

Load these for detailed patterns:

• 1-ingestion-patterns.md - Auto Loader, Kafka, Event Hub, Kinesis, file formats
• 2-streaming-patterns.md - Deduplication, windowing, stateful operations, joins
• 3-scd-patterns.md - Querying SCD Type 2 history tables, temporal joins
• 4-performance-tuning.md - Liquid Clustering, optimization, state management
• 5-python-api.md - Modern dp API vs legacy dlt API comparison
• 6-dlt-migration.md - Migrating existing DLT pipelines to SDP
• 7-advanced-configuration.md - extra_settings parameter reference and examples
• 8-project-initialization.md - Using databricks pipelines init, Asset Bundles, language detection, and migration guides

Best Practices (2025)

Project Structure

• Default to databricks pipelines init for new projects (creates Asset Bundle)
• Use Asset Bundles for multi-environment deployments (dev/staging/prod)
• Manual structure only for quick prototypes or legacy migration
• Medallion architecture: Two approaches work with Asset Bundles:
  • Flat structure (template default): bronze_*.sql, silver_*.sql, gold_*.sql in transformations/
  • Subdirectories: transformations/bronze/, transformations/silver/, transformations/gold/
  • Both work with the transformations/** glob pattern - choose based on team preference
• See 8-project-initialization.md for project setup details

Modern Defaults

• CLUSTER BY (Liquid Clustering), not PARTITION BY - see 4-performance-tuning.md
• Raw .sql/.py files, not notebooks
• Serverless compute ONLY - Do not use classic clusters unless explicitly required
• Unity Catalog (required for serverless)
• read_files() for cloud storage ingestion - see 1-ingestion-patterns.md

Reading Tables in Python

Modern SDP Best Practice:

• Use spark.read.table() for batch reads
• Use spark.readStream.table() for streaming reads
• Don't use dp.read() or dp.read_stream() (old syntax, no longer documented)
• Don't use dlt.read() or dlt.read_stream() (legacy DLT API)

Key Point: SDP automatically tracks table dependencies from standard Spark DataFrame operations. No special read APIs are needed.

Three-Tier Identifier Resolution

SDP supports three levels of table name qualification:

Option 1: Unqualified Names (Recommended for Pipeline Tables)

Best practice for tables within the same pipeline. SDP resolves unqualified names to the pipeline's configured target catalog and schema. This makes code portable across environments (dev/prod).

@dp.table(name="silver_clean")
def silver_clean():
    # Reads from pipeline's target catalog/schema (e.g., dev_catalog.dev_schema.bronze_raw)
    return (
        spark.read.table("bronze_raw")
        .filter(F.col("valid") == True)
    )

@dp.table(name="silver_events")
def silver_events():
    # Streaming read from same pipeline's bronze_events table
    return (
        spark.readStream.table("bronze_events")
        .withColumn("processed_at", F.current_timestamp())
    )

Option 2: Pipeline Parameters (For External Sources)

Use spark.conf.get() to parameterize external catalog/schema references. Define parameters in pipeline configuration, then reference them at the module level.

from pyspark import pipelines as dp
from pyspark.sql import functions as F

# Get parameterized values at module level (evaluated once at pipeline start)
source_catalog = spark.conf.get("source_catalog")
source_schema = spark.conf.get("source_schema", "sales")  # with default

@dp.table(name="transaction_summary")
def transaction_summary():
    return (
        spark.read.table(f"{source_catalog}.{source_schema}.transactions")
        .groupBy("account_id")
        .agg(
            F.count("txn_id").alias("txn_count"),
            F.sum("txn_amount").alias("account_revenue")
        )
    )

Configure parameters in pipeline settings:

• Asset Bundles: Add to pipeline.yml under configuration:
• Manual/MCP: Pass via extra_settings.configuration dict

# In resources/my_pipeline.pipeline.yml
configuration:
  source_catalog: "shared_catalog"
  source_schema: "sales"

Option 3: Fully-Qualified Names (For Fixed External References)

Use when referencing specific external tables that don't change across environments:

@dp.table(name="enriched_orders")
def enriched_orders():
    # Pipeline-internal table (unqualified)
    orders = spark.read.table("bronze_orders")

    # External reference table (fully-qualified)
    products = spark.read.table("shared_catalog.reference.products")

    return orders.join(products, "product_id")

Choosing the Right Approach

Common Issues

For detailed errors, the result["message"] from create_or_update_pipeline includes suggested next steps. Use get_pipeline_events(pipeline_id=...) for full stack traces.

Advanced Pipeline Configuration

For advanced configuration options (development mode, continuous pipelines, custom clusters, notifications, Python dependencies, etc.), see 7-advanced-configuration.md.

Platform Constraints

Serverless Pipeline Requirements (Default)

Serverless Limitations (When Classic Clusters Required)

General Constraints

Default to serverless unless user explicitly requires R, RDD APIs, or JAR libraries.