Home › Key concepts › ETL vs ELT

Key concept · Crawl (beginner) · the same on every platform

ETL vs ELT

Two orders for the same three steps: extract, transform, load.

What it is

ETL means you extract data from a source, transform it on the way, then load the clean result into the warehouse. ELT means you extract and load the raw data first, then transform it inside the warehouse with SQL. Cheap storage and powerful warehouses made ELT the usual choice today.

ETLExtractfrom sourceTransformoutside thewarehouseLoadclean dataonlyELTExtractfrom sourceLoadraw datakeptTransformwith SQL insidethe warehouse
Same three steps, different order. ELT keeps the raw copy, so you can re-run a fixed transformation.

Why it matters

With ELT you always keep the raw data, so you can fix a transformation and re-run it without going back to the source.

An example

A shop's orders are copied raw into the lake every hour (extract, load). A SQL model then cleans them into a daily sales table (transform).

The common mistake

Transforming before you land the raw copy, then discovering a bug and having nothing to re-run from.

Where you meet it on each platform

The idea stays the same; only the product name changes.

JobAWSAzureGoogle CloudDatabricksOpen source
Ingestioncopy data in from apps, databases and APIsAWS DMS / AppFlowData Factory copy / Fabric mirroringDatastreamLakeflow Connect / Auto Loaderdlt or Airbyte
SQL modellingturn raw tables into clean, tested, documented onesdbt on Redshift or Athenadbt on FabricDataform or dbt on BigQueryLakeflow Declarative Pipelines or dbtdbt with DuckDB

Related ideas

Where it shows up

Part of the Crawl level on these paths: Data Engineering (moving data).