Purview Deep Dive: Classification, Lineage & Power BI

Master Purview with advanced classification, lineage tracking, and Power BI integration

Purview Capabilities

Architecture Diagram

┌─────────────────────────────────────────────────────────────────────┐
│                    PURVIEW CAPABILITIES                              │
├─────────────────────────────────────────────────────────────────────┤
│                                                                     │
│  AUTOMATED DISCOVERY & SCANNING                                     │
│  ┌─────────────────────────────────────────────────────────────┐   │
│  │ • Connect to 100+ data sources                               │   │
│  │ • Scheduled scanning (daily/weekly)                          │   │
│  │ • Incremental scanning (changed data only)                   │   │
│  │ • Custom scan rulesets                                       │   │
│  └─────────────────────────────────────────────────────────────┘   │
│                                                                     │
│  CLASSIFICATION                                                      │
│  ┌─────────────────────────────────────────────────────────────┐   │
│  │ • 100+ built-in classifiers (PII, financial)                 │   │
│  │ • Custom classifiers (regex, keyword list)                   │   │
│  │ • Auto-labeling with sensitivity labels                      │   │
│  │ • Column-level classification                                 │   │
│  └─────────────────────────────────────────────────────────────┘   │
│                                                                     │
│  LINEAGE                                                            │
│  ┌─────────────────────────────────────────────────────────────┐   │
│  │ • End-to-end lineage across Azure services                   │   │
│  │ • ADF pipeline lineage                                       │   │
│  │ • Databricks notebook lineage                                │   │
│  │ • Synapse SQL lineage                                        │   │
│  │ • Power BI dataset lineage                                   │   │
│  └─────────────────────────────────────────────────────────────┘   │
│                                                                     │
│  DATA MAP & CATALOG                                                 │
│  ┌─────────────────────────────────────────────────────────────┐   │
│  │ • Unified metadata repository                                │   │
│  │ • Business glossary                                          │   │
│  │ • Search and discovery                                       │   │
│  │ • Impact analysis                                            │   │
│  └─────────────────────────────────────────────────────────────┘   │
└─────────────────────────────────────────────────────────────────────┘

Custom Classification

from azure.purview.datamap import PurviewDataMapClient
from azure.identity import DefaultAzureCredential

credential = DefaultAzureCredential()
client = PurviewDataMapClient(credential=credential, account_name="purview-prod")

# Create custom classifier
client.classification.create_classification_rule(
    rule={
        "name": "CustomerCodeClassifier",
        "description": "Detects internal customer codes",
        "ruleType": "Regex",
        "pattern": r"CUST-\d{6}",
        "columnNamePatterns": ["customer_id", "cust_code"],
        "minPrecision": 0.8
    }
)

# Apply custom classifier
client.classification.classify_asset(
    asset_type="azure_datalake_gen2_path",
    asset_guid="asset-guid",
    classifiers=["CustomerCodeClassifier"]
)

Lineage Tracking

# Get lineage for a data asset
lineage = client.lineage.get_lineage(
    entity_guid="asset-guid",
    direction="Both"
)

# Visualize lineage graph
for edge in lineage.relations:
    print(f"{edge.source_entity.name} --> {edge.target_entity.name}")
    print(f"  Type: {edge.relationship_type}")

Power BI Integration

{
  "scanName": "powerbi-workspace-scan",
  "dataSource": {
    "type": "PowerBI",
    "properties": {
      "tenantId": "tenant-id",
      "workspaceIds": ["workspace-id-1", "workspace-id-2"]
    }
  },
  "scanRuleset": {
    "type": "System",
    "name": "PowerBI"
  }
}

ℹ️

Pro Tip: Use Purview's lineage to trace data from source to Power BI report. This enables impact analysis when source schemas change and helps with data trust assessment.

Interview Questions

Q1: How do you implement end-to-end lineage in Purview? A: Enable integrations with ADF, Databricks, Synapse, and Power BI. Configure scanning schedules. Use Purview SDK to register custom lineage. Link glossary terms to assets for business context.

Q2: What are the best practices for Purview scanning? A: 1) Scan by domain/collection, 2) Use appropriate scan rulesets, 3) Schedule incremental scans, 4) Monitor scan status, 5) Review and approve classification results, 6) Use custom classifiers for domain-specific data.

Q3: How does Purview support data governance? A: Provides automated discovery, classification, lineage tracking, business glossary, access policies, and compliance reporting—enabling organizations to understand, manage, and protect their data assets.