Once the data is extracted or delivered, it is generally staged into temporary storage for further processing. It is generally a good idea to keep data extraction/delivery storage separate from staging storage, although there are instances where this won't be necessary.
The staging area cleanly separates the following two aspects of the data ingestion process:
- Data that has been extracted or delivered
- New data that has to be processed
Once the data is staged completely, when we reach the further processing steps, such as cleaning, we do not have to be concerned about new data arriving. We can think of staged data as something that, once created, never changes and is immutable. This means that, to an already staged piece of data, no more data can be added, deleted, or modified. This is a very important data property that is necessary for reliable data processing...