Skip to content

Teradata to Azure Migration — Complete Package¶

Audience: Enterprise data teams running Teradata (on-prem appliances or VantageCloud) evaluating or executing a migration to Azure (Synapse, Databricks, Fabric). This hub page links every artifact in the package and provides a decision matrix for choosing your migration path.

Quick-start decision matrix¶

Your situation	Recommended starting point	Primary Azure target
Executive asking "should we migrate?"	Why Azure over Teradata then TCO Analysis	N/A — strategic read
Need a business case with numbers	TCO Analysis	N/A — financial read
Architect mapping Teradata features	Feature Mapping	Depends on feature
DBA converting SQL scripts	SQL Migration	Synapse / Databricks
ETL team replacing TPT/BTEQ	Data Migration, Tutorial — TPT to ADF	ADF + ADLS + Delta
Data engineer converting BTEQ to dbt	Tutorial — BTEQ to dbt	Databricks / Fabric
Platform team mapping TASM workloads	Workload Migration	Synapse / Databricks
Security / compliance team	Security Migration	Entra + Purview + Monitor
Performance validation team	Benchmarks	Synapse / Databricks
Migration lead planning phases	Best Practices then this page (Gantt)	All

Package contents¶

Strategic and financial¶

Document	Lines	Purpose
Why Azure over Teradata	~400	Strategic rationale, market position, honest pros/cons
TCO Analysis	~350	5-year cost model: on-prem appliance vs Azure vs VantageCloud

Technical migration guides¶

Document	Lines	Purpose
Feature Mapping (Complete)	~400	40+ Teradata features mapped to Azure equivalents
SQL Migration	~450	25+ SQL conversion patterns with before/after examples
Data Migration	~400	TPT export, ADF ingestion, validation frameworks
Workload Migration	~350	TASM classes to Azure workload management
Security Migration	~300	Access logging, RLS, roles, encryption mapping

Hands-on tutorials¶

Document	Lines	Purpose
Tutorial — BTEQ to dbt	~350	End-to-end: convert a BTEQ script to a tested dbt model
Tutorial — TPT to ADF	~350	End-to-end: replace TPT pipeline with ADF + ADLS + dbt

Validation and planning¶

Document	Lines	Purpose
Benchmarks	~300	Performance, concurrency, cost-per-query comparisons
Best Practices	~300	Schema assessment, phased cutover, common pitfalls

Original guide¶

Document	Lines	Purpose
Teradata Migration Overview	~205	Foundational guide covering architecture, 5 phases, cost, pitfalls

Typical migration timeline (12-24 months)¶

The Gantt chart below shows a representative timeline for a mid-to-large Teradata estate (50-500 TB, 2,000-10,000 tables). Adjust durations to your specific environment.

gantt
    title Teradata → Azure Migration (18-month reference)
    dateFormat  YYYY-MM
    axisFormat  %b %Y

    section Phase 1 — Assessment
    Inventory & discovery           :a1, 2025-01, 2025-02
    Workload profiling (TASM)       :a2, 2025-02, 2025-03
    Tool evaluation (SAMA, sqlglot) :a3, 2025-02, 2025-03
    Migration tier classification   :a4, 2025-03, 2025-04
    Business case & TCO             :a5, 2025-03, 2025-04

    section Phase 2 — Design
    Target architecture design      :d1, 2025-04, 2025-05
    Schema mapping (PI → partition) :d2, 2025-05, 2025-06
    Security model design           :d3, 2025-05, 2025-06
    Network & ExpressRoute plan     :d4, 2025-05, 2025-06

    section Phase 3 — Migration
    Azure landing zone deployment   :m1, 2025-06, 2025-07
    Tier-D decommission             :m2, 2025-06, 2025-08
    Tier-A bulk SQL translation     :m3, 2025-07, 2025-10
    Tier-B manual refactor          :m4, 2025-08, 2026-01
    Tier-C architectural rework     :m5, 2025-10, 2026-03
    Data bulk load (TPT → ADLS)     :m6, 2025-07, 2025-09
    CDC / continuous sync           :m7, 2025-09, 2026-04
    dbt model development           :m8, 2025-08, 2026-02

    section Phase 4 — Cutover
    Parallel run (per schema)       :c1, 2026-01, 2026-04
    BI re-validation                :c2, 2026-02, 2026-05
    Consumer repointing             :c3, 2026-03, 2026-05
    Teradata read-only              :c4, 2026-04, 2026-05

    section Phase 5 — Decommission
    Final extract & archive         :x1, 2026-05, 2026-06
    License termination             :x2, 2026-06, 2026-06
    Hardware decommission           :x3, 2026-06, 2026-07

Migration tiers (from original guide)¶

Every workload should be classified before migration begins.

Tier	Description	Typical percentage	Action
A Direct migrate	Pure SQL, standard ANSI features	30-40%	Automated translation via SAMA / sqlglot
B Refactor required	Teradata-specific SQL (QUALIFY, RECURSIVE, MERGE)	20-30%	Manual rewrite to Spark SQL or T-SQL
C Architectural rework	TASM-dependent, custom Java UDFs, QueryGrid	10-20%	Redesign in dbt + Databricks / Synapse
D Decommission	Zombie workloads, unused tables, dead ETL	20-40%	Archive output, delete

Recommended reading order¶

For executives (2-hour read):

Why Azure over Teradata
TCO Analysis
Benchmarks — executive summary section

For architects (4-hour read):

For hands-on engineers (full day):

Teradata Migration Overview — the original 205-line guide this package expands
Migrations — Hadoop / Hive — similar phased pattern
Migrations — Snowflake — sister cloud DW migration
Reference Architecture — Fabric vs Synapse vs Databricks
Patterns — Power BI & Fabric Roadmap
Microsoft SAMA: https://aka.ms/sama
Azure for Teradata customers: https://learn.microsoft.com/azure/architecture/databases/idea/teradata-migration
sqlglot: https://github.com/tobymao/sqlglot

Maintainers: csa-inabox core team Last updated: 2026-04-30