Extract, Load and Transform (ELT) is a data manipulation process in database usage, especially in data warehousing. Different from the traditional ETL (Extract, Transform, Load) mode, in ELT, data is extracted, loaded into the database and then is transformed where it sits in the database, prior to use. This data is migrated in bulk according to the data set and the transformation process occurs after the data has been loaded into the targeted DBMS in its raw format. This way, less stress is placed on the network and larger throughput is gained.
However, the ELT mode is certainly not optimal for all situations, for example,
As SQL is less powerful than Java, the scope of available data transformations is limited.
ELT requires users that have high proficiency in SQL tuning and DBMS tuning.
Using ELT with Talend Studio, you cannot pass or reject one single row of data as you can do in ETL. For more information about row rejection, see Row connection.
Based on the advantages and disadvantages of ELT, the SQL templates are designed as the ELT facilitation requires.