![]() Managing PII (personally identifiable information) is an example of the ETL use case. If your goal is analytics, and you are using an OLAP database, you will almost always choose “ELT,” unless you have a specific use case where specific data needs to be cleansed or transformed before the load. If you’re building an application on an OLTP database, you will almost always choose “ETL,” as you can control changes to the data at the row level. In this approach, the goal is to enable more users to apply their own business logic by replicating data from the source in the same format and then applying the transformation, post-load. This approach improves the speed and ease of making changes to the transformation logic, as this is often done in the database using SQL or Python and reduces the need for developers. In the last decade, with the proliferation of APIs and the need to access data from cloud-based applications, “ELT” was born. ![]() Since few people in an organization have the ability to read or modify the application, this makes the business logic in the “ETL” process harder to understand and also harder to modify. In the older and more traditional “ETL” approach, the transformation is done after the data is extracted into a temporary location (a "staging area") and before the data is loaded. The full ETL process almost always requires a developer, and often requires a custom application to handle each step.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |