ETL Process:
ETL, an acronym for ‘Extraction, Transformation and Loading’ is a collection of processes associated with extracting the source data, transforming that data and finally loading that data into a data warehouse. Before loading the required into data warehouse, it should be transformed in order to meet the needs of the data warehouse. This transformation involves several processes like data cleansing, data profiling, data type conversion, validating for referential integrity, performing aggregation if needed, denormalization and normalization.
Data Warehouse:
Data warehouse is a centralized repository where all the information for analysis is kept in an organization. This is the data collected from variant sources for the purpose of analytical processing and reporting. This data is non-volatile and a data warehouse is built on a dimensional data model. From this data warehouse, data can be extracted for reporting needs with the help of query tools or many data marts can be built based on subject area requirements.
Data Mart:
Data Mart is subject oriented, basically a sub-set of data warehouse, built for the purpose of analyzing a particular line of business or department. It holds the data specific to a particular subject area like sales, purchase etc. Data marts can be of derived from a data warehouse or built for the sole purpose of BI directly from the source and like data warehouse, data marts are also constructed from dimensional data models.
