A key component to Business Intelligence (BI) architecture is the Extract, Translate and Load (ETL) process. The ETL process is typically scheduled on a nightly basis and is responsible for moving data from one or more source systems into a data warehouse. In early data warehouses, ETL was typically performed by custom-developed programs and scripts. Over the past twenty years, ETL development has matured into configurable tools. Two common tools used with Oracle Business Intelligence include Informatica Power Center and(ODI). While these tools have similar features, there are considerations that can be used to identify the correct tool for a particular Business Intelligence environment.
What does ETL or ELT do? Both ETL and ELT extract the data from one or many source systems (majorly transactional systems), then load it to a target system (majorly data warehouse systems). There will be data transformations performed on the source data to fit the target requirements. These transformations are done between the extract and load process or after the load process based on the environment. In order to keep the source and target system highly available, the data is extracted as fast as possible from the source with minimum to no impact on the source and loaded into the target with minimal impact on the target to maintain system availability. As target database systems have been evolving, they have become more efficient in processing data transformations. This has also evolved into the option to have transformations performed after loading the data into the target.
ETL stands for Extract, Transform and Load. ETL is typically performed by an external tool such as Informatica, DataStage or by PLSQL procedures. When tools like Informatica are used, most of the data transformation is done on the ETL server, which made the ETL server to be well built to support the data transformation process. The advantage of this was less impact on the source and target systems.
ELT stands for extract, load, and transform. The majority of the data transformation is done after the data is loaded into the target database. This will reduce the load on the ETL server and in turn the cost of the server. Since most of the transformation is done on the database, it is also faster, which in turn increases the performance and decreases the duration of the data movement window. This utilizes the capabilities of the target database to increase the performance.
We also provides the best ETL Testing Online Training will assist you to start the flying career in the world of ETL Testing.