ELT
ELI5 — The Vibe Check
ELT is ETL's modern cousin. Instead of transforming data before loading it, you dump the raw data into your warehouse first, then use the warehouse's beefy compute to transform it there. Modern warehouses like Snowflake and BigQuery are so powerful that transforming inside them is often faster than doing it externally.
Real Talk
ELT (Extract, Load, Transform) loads raw data directly into the target system before applying transformations using the target's processing power. This approach leverages the scalability of modern cloud data warehouses and preserves raw data for future use cases. dbt has become the standard tool for the Transform step in ELT pipelines.
When You'll Hear This
"We switched from ETL to ELT because BigQuery's compute is faster than our ETL server." / "dbt handles the T in our ELT pipeline."
Related Terms
Data Lakehouse
A data lakehouse is what you get when a data lake and a data warehouse have a baby.
Data Warehouse
A data warehouse is where all your company's data goes to be analyzed.
ETL
ETL stands for Extract, Transform, Load. You extract data from sources, transform it (clean, reshape, calculate), then load it into your warehouse.
OLAP
OLAP is all about analyzing huge amounts of data to answer business questions. 'What were total sales by region last quarter?' That's an OLAP query.