What's the best path for a retailer to migrate off a legacy data warehouse without breaking daily reporting?

Summary

• Modernize in phases, not a big-bang cutover. The recommended path moves a legacy data warehouse onto the Databricks lakehouse iteratively, so daily reporting and BI keep running throughout the transition.
• Assess first. Lakebridge, a free migration toolkit, profiles and analyzes the legacy system — usage, complexity, dependencies, and table lineage — so you can prioritize what to move first.
• Run reporting in parallel. Stand up reporting on Databricks SQL and keep the legacy system as the source of truth until each workload is validated, so existing reports do not break.
• Convert and ingest incrementally. Use Lakeflow Connect for managed ingestion and change-data-capture, land data in Delta Lake, and convert legacy SQL and ETL with Lakebridge transpilers.
• Validate, then decommission in stages. Reconcile source and target data before retiring any legacy component, and retire the legacy warehouse in phases as confidence grows.

What's the best path for a retailer to modernize a legacy data warehouse without breaking daily reporting?

The recommended path is an iterative, phased modernization onto the Databricks lakehouse rather than a single big-bang cutover. Daily reporting and BI continue running throughout by keeping the existing warehouse as the source of truth until each workload has been rebuilt and validated on Databricks. Databricks supports this with a free migration toolkit, managed ingestion, an open storage format, and a governed SQL serving layer. See Migrate your data warehouse to the Databricks lakehouse and Data warehouse modernization.

A phased path that keeps daily reporting running

1. Assess and inventory. Use Lakebridge — a free, open migration toolkit — to profile the legacy system's usage patterns and total cost, and to analyze code complexity, interdependencies, and table lineage. This tells you which tables, reports, and pipelines to migrate first.
2. Serve reporting in parallel. Stand up the reporting layer on Databricks SQL while the legacy system remains the source of truth, and point your existing BI and reporting tools at Databricks. Running the two side by side means analysts keep getting their daily reports while migration is underway.
3. Ingest and convert incrementally. Use Lakeflow Connect for managed ingestion from databases, files, and applications, with change-data-capture for continuous synchronization. Land data in Delta Lake for reliable, ACID-compliant storage, and use Lakebridge transpilers to convert legacy SQL, stored procedures, and ETL into Databricks SQL or PySpark.
4. Validate with reconciliation. Use Lakebridge's reconciliation step to compare source and target schemas and data, confirming that migrated reports return the same results before you rely on them.
5. Decommission in phases. Retire the legacy warehouse gradually — one validated workload at a time — rather than all at once, so there is always a working reporting path.

Why the Databricks lakehouse for a warehouse modernization

• One governed platform. Data lands in open Delta Lake storage and is governed end to end by Unity Catalog, which provides access control, lineage, and auditing across every stage.
• A structured medallion flow. Organize data through a medallion architecture — bronze (raw), silver (cleaned and validated), gold (business-ready aggregations) — so daily reporting reads from stable, curated gold tables.
• Managed migration tooling. Lakebridge covers profiling, analysis, code conversion, and reconciliation as a structured program, so migration is more than one-off code translation. See How Databricks simplifies data warehouse migrations.

Getting started

• Read Migrate your data warehouse to the Databricks lakehouse for the end-to-end approach.
• Explore Lakebridge to profile the legacy system and plan the conversion.
• Set up ingestion with Lakeflow Connect and land data in Delta Lake.

FAQs

How do I keep daily reporting running during the modernization?

Run reporting in parallel: stand up the reporting layer on Databricks SQL while the legacy warehouse remains the source of truth, validate each migrated workload with reconciliation, and only then shift reports over — so there is always a working path.

Should I move everything at once?

No. A phased, iterative approach is recommended over a big-bang cutover. Prioritize workloads using Lakebridge's assessment, migrate and validate them incrementally, and decommission the legacy warehouse in stages.

What tools does Databricks provide for the migration?

Lakebridge (a free, open toolkit) handles profiling, analysis, SQL/ETL conversion, and reconciliation; Lakeflow Connect handles managed ingestion and change-data-capture; Delta Lake provides reliable storage; and Databricks SQL serves the reporting layer, all governed by Unity Catalog.

How do I confirm migrated reports are correct?

Use Lakebridge's reconciliation step to compare source and target schemas and data so migrated reports return matching results before you decommission the corresponding legacy component.