ELT stands for “Extract, Load, Transform”. ELT is a data process used to replicate data from a source to a destination database, being an evolution of the more commonly known ETL (extract, transform and load) as it makes the data replication process much less complex, a since the transformation step is performed after the data is at the destination.
The ELT steps:
Extract: Copy data from a source source.
Load: Replicate this data to the target database.
Transform: Model the data to structure it, facilitate queries, cross with other sources and analyze it.
Why is it different from ETL?
The difference between ETL and ELT is more significant than just a change in letter order. In ETL, where data transformation is done before writing to the final destination, the data replication process is more complex to be developed as it requires a developer with specific knowledge to create it and change it whenever there is a change in the data structure/pattern or the need for some change in the transformation requested by a business analyst.
Meanwhile, in the ELT, the data loading process can be done without any complexity using tools like Kondado, and the data modeling and transformation part can be done by the business analysts themselves, who will be able to make changes according to their needs and without depend on a developer for this.
Naturally, the ELT requires more resources from the data target, as this is now where data transformations are performed. However, today the power of data targets available in the cloud makes ELT a simpler and more agile option to ETL and the perfect choice for the most dynamic data areas.
