In the past, processing large data sets impacted an organization’s computing power and so these processes were performed in batches during off-hours. There are four primary types of ETL tools:īatch Processing: Traditionally, on-premises batch processing was the primary ETL process. See a side-by-side review of 10 key areas in the ETL vs ELT Comparison Matrix. This is why the ELT process is more appropriate for larger, structured and unstructured data sets and when timeliness is important. But if there is not sufficient processing power in the cloud solution, transformation can slow down the querying and analysis processes. This means that the ELT process takes less time. In the ELT process, data transformation is performed on an as-needed basis within the target system. This is why this process is appropriate for small data sets which require complex transformations. The benefit is that analysis can take place immediately once the data is loaded. The entire data set must be transformed before loading, so transforming large data sets can take a lot of time up front. In the ETL process, transformation is performed in a staging area outside of the data warehouse and before loading it into the data warehouse. For larger data sets and when timeliness is important, the ELT process is more appropriate.

The ETL process is most appropriate for small data sets which require complex transformations. Many organizations use both processes to cover their wide range of data pipeline needs.

The key difference between the two processes is when the transformation of the data occurs.

Extract load transform tutorial