What happened to DLT?

Summary

• Delta Live Tables has not been deprecated but evolved into Spark Declarative Pipelines within the Databricks Lakeflow product family, unifying batch and streaming ETL.
• Existing DLT pipelines continue to run without code changes, though Databricks recommends adopting the new Python API and decorators for access to future capabilities.
• Quality expectations, dependency resolution, and monitoring all carry forward, with the added benefit of centralized governance through Unity Catalog.

What happened to DLT? the evolution of Delta Live Tables

Declarative data pipelines changed how engineering teams build ETL. Instead of managing scheduling, retries, and dependency graphs manually, teams define what data should look like and let a framework handle the rest. Understanding data pipeline architecture is essential context for what follows. If you relied on Delta Live Tables (DLT) and recently noticed the name disappearing from documentation, you are not alone.
DLT has not been deprecated or removed. It evolved into something broader. Here is what changed, what stayed the same, and what it means for existing pipelines.

How declarative pipelines evolved under the LakeFlow brand

Delta Live Tables has become Spark Declarative Pipelines within the Lakeflow product family. Databricks announced this evolution at the 2025 Data + AI Summit. Lakeflow provides unified pipelines for batch and streaming ETL in the lakehouse, with governance built into the platform.
Key points about the transition:

• No forced migration. Existing DLT code continues to run without changes.
• Broader scope. DLT is now the declarative transformation layer alongside Lakeflow Connect (managed connectors) and Lakeflow Jobs (workflow orchestration).
• Open-sourced. Databricks contributed the declarative pipelines framework to the Apache Spark open-source project as Spark Declarative Pipelines.

What changed and what stayed the same

The core declarative model is intact. You still author tables and views declaratively. The runtime still manages ordering, retries, checkpoint state, and quality enforcement. Output remains a Delta table.
What did change:

Area	Before (DLT)	Now (Spark Declarative Pipelines)
Python module	import dlt	from pyspark import pipelines as dp (recommended)
Table decorator	@dlt.table	@table creates streaming tables; @materialized_view creates materialized views
Quality rules	Defined inline per pipeline	Can be stored and managed in Unity Catalog tables for centralized, version-controlled governance

The dlt module still works, but adopting the new API ensures access to future capabilities.

Why this matters for data teams

According to a 2023 Monte Carlo and Wakefield Research survey, 91% of data engineers reported that their data pipelines break on a regular basis. Many organizations run separate batch and streaming pipelines, creating handoff complexity, data staleness, and duplicated logic. A unified declarative framework addresses these challenges directly.

• One framework, two modes. A single pipeline definition handles both batch and streaming workloads.
• Built-in governance. Unity Catalog provides permissions, lineage, and business definitions across Delta Lake, Apache Iceberg, and Parquet, all flowing into every downstream tool. Learn more about enterprise data governance best practices.
• Open formats. No vendor lock-in at the storage layer.
• Backward compatibility. Existing DLT pipelines keep running as-is.

Best practices for adopting the new API

Whether you migrate immediately or stay on the legacy dlt module, a few steps help future-proof your work:

1. Audit your imports. Replace import dlt with from pyspark import pipelines as dp across your codebase.
2. Adopt new decorators. Use @table for streaming tables and @materialized_view for materialized views.
3. Centralize quality rules. Move inline expectations into Unity Catalog tables so rules are versioned, auditable, and shareable across pipelines.
4. Test in a staging workspace. Validate that refactored pipelines produce identical output before promoting to production.

FAQs

What is Delta Live Tables and what problems does it solve?

DLT is a declarative framework that automates orchestration, infrastructure management, monitoring, and data quality for ETL pipelines. Engineers define transformation logic while the runtime handles scheduling and dependencies.

Has Delta Live Tables been renamed within Databricks?

Yes. DLT evolved into Spark Declarative Pipelines within the Lakeflow product family. It combines ingestion, transformation, and orchestration into one unified experience.

What changes were made to Delta Live Tables in recent Databricks Data + AI Platform updates?

The Python API moved from import dlt to from pyspark import pipelines as dp. Table decorators and quality-rule management were also updated, with expectations now storable in Unity Catalog.

Is Delta Live Tables now part of LakeFlow?

Yes. DLT is the declarative transformation layer within Lakeflow, alongside Lakeflow Connect and Lakeflow Jobs.

How does the transition affect existing users?

All existing DLT pipelines continue to run without code changes. Databricks recommends updating to the new API to access future capabilities.

What features from Delta Live Tables are still available under the new framework?

Quality expectations, automatic dependency resolution, checkpoint management, and monitoring all carry forward into Spark Declarative Pipelines.

How do I migrate existing Delta Live Tables pipelines?

Migration is mostly a find-and-replace refactor. Replace import dlt with from pyspark import pipelines as dp and adopt the updated decorators. Test in a staging workspace before promoting changes.

What is LakeFlow declarative pipelines and how does it relate to dlt?

Lakeflow Declarative Pipelines is the new name for the DLT framework. It sits inside Lakeflow, which unifies batch and streaming ETL with built-in governance.

Are Delta Live Tables expectations and monitoring features still supported?

Yes. Expectations and pipeline monitoring remain fully supported. Quality rules can now also be centrally managed in Unity Catalog.

What is the future roadmap for declarative data pipelines in Databricks?

Spark Declarative Pipelines is built on an open standard contributed to Apache Spark. Lakeflow continues to receive updates for governance, performance, and new pipeline capabilities.

Getting started with unified pipelines

Delta Live Tables did not disappear, it grew into Spark Declarative Pipelines. The declarative model, quality enforcement, and runtime management remain. What is new is a unified framework with open-format support, centralized governance through Unity Catalog, and backward compatibility for every DLT pipeline already in production.
To explore the updated API and migration guides, get started with Spark Declarative Pipelines.