ELTtüüpi

ELTtüüpi refers to a data integration and processing approach where data is Extracted, Loaded, and then Transformed within the target system. It is an alternative to the traditional ETL model, where transformation occurs before loading. The term is used in Estonian to describe the type of workflow rather than a specific technology.

In an ELT-type pipeline, raw data is ingested into a data warehouse or data lake in its native form, and the heavy lifting of data transformation is performed inside the target environment. This leverages the compute power and scalability of modern cloud-based warehouses, typically using SQL or specialized ELT tooling. Transformation steps may include cleaning, joining, aggregating, and shaping data into analytic models or curated tables and views.

Key advantages of ELT-type architectures include improved scalability for large and diverse data sets, reduced data movement, and faster ingestion, since data can be loaded quickly in its raw form and transformed on-demand. They also support flexible data modeling, such as schema-on-read or iterative modeling with versions of data marts.

Challenges involve governance and quality control, as raw data resides in the target system before cleansing. Performance is dependent on the efficiency of the warehouse or lake, and complex transformations can require careful optimization and monitoring. Strong metadata management and testing are essential to maintain reliable analytics.

ELTtüüpi is commonly used with cloud data warehouses like Snowflake, BigQuery, and Redshift, often in combination with SQL-based transformation tools such as dbt. It is favored for modern data architectures that prioritize scalability, speed of ingestion, and flexible data modeling.