🔄ELT

NEW

Also known as: Extract-Load-Transform

A data integration pattern that loads raw data first, then performs transformations within the target system.

DESCRIPTION

ELT leverages the compute power of modern cloud warehouses and lakehouses by loading raw data and transforming it in-place using SQL or engines. It simplifies ingestion, preserves raw lineage, and scales via pushdown execution. dbt and similar tools orchestrate models and tests in-warehouse. ELT suits analytics where transformation logic evolves rapidly. Robust governance ensures raw and transformed layers remain controlled and auditable.

ALTERNATIVE DEFINITIONS

Snowflake

ELT (extract, load, transform) is a data integration process that extracts data from sources, loads it into a target system, and then transforms it within that system.

Emphasizes in-platform transformation and modern cloud practices.

Source

Google Cloud

Extract, load, transform (ELT) moves data into a target system first and then transforms it there, taking advantage of the target’s compute and scalability.

Highlights pushdown and scalability in the destination platform.

Source

RELATED TERMS

Data Warehouse Data Lakehouse dbt SQL Orchestration Change Data Capture Pipelines Materialized View Query Pushdown

COMPETING TERMS

Extract-Transform-Load

KNOWLEDGE AREAS

Data Engineering Data Integration & Interoperability Data Warehousing & BI Cloud

TARGET AUDIENCE

Data Engineer Analytics Engineer Data Architect BI Specialist

EXTERNAL SOURCES

• Wikipedia - Extract, transform, load#ELT
• Google Cloud - Google Cloud overview of ELT
• Snowflake - Snowflake guide to ELT
• Microsoft - Azure guidance on ETL and ELT

TOP QUESTIONS

Why choose ELT over ETL?

ELT exploits target platform compute for scalable transformations, reduces pre-processing, and preserves raw data for re-modeling. It’s ideal for cloud warehouses and lakehouses.

Category: Decision

What tools enable ELT?

dbt, native SQL engines, stored procedures, materialized views, and task schedulers enable ELT. Cloud ingestion tools handle extract and load steps.

Category: Tooling

How does ELT affect data governance?

ELT necessitates clear layer separation (raw, staging, curated), robust testing, and lineage. Role-based access controls protect sensitive raw data.

Category: Governance

Is ELT suitable for real-time workloads?

Micro-batch ELT can support near-real-time use cases, but true sub-second latency often favors streaming engines or CDC pipelines.

Category: Scope

What are common mistakes with ELT?

Uncontrolled sprawl of Ad-hoc SQL, missing tests and documentation, and costly, unoptimized transformations are typical pitfalls.

Category: Mistakes

How does ELT impact cost?

Costs shift toward warehouse compute; efficient SQL, scheduling, and incremental models help control spend.

Category: Cost

Can ELT and ETL coexist?

Yes. Many architectures use ETL for sensitive pre-cleaning and ELT for scalable in-warehouse modeling.

Category: Practice

What is a good starting point for ELT?

Begin with a small domain, define raw and curated layers, adopt dbt for modeling and tests, and automate CI/CD.

Category: Getting Started

How does ELT support data reusability?

By storing raw data and building modular models, ELT enables reuse for new analytics without re-extraction.

Category: Benefits