What is SCD?

SCD is a short form used for ‘Slowly Changing Dimensions.’ This is a concept of how dimensions or attributes in a data warehouse change with time. One of the fundamental reasons we consider SCDs while discussing a slowly changing dimension is to handle historical data or data that change very slowly with customer information, product details, or geographic data. The core objective of managing slowly changing dimensions is focused on the maintenance of historical accuracy and integrity of data.

SCDs track these changes over time since data analysis is performed. The following are the three major types of SCDs:

Type 1 SCD: This type overwrites the old value with the new one. This is considered the simplest form of SCD, but in practicality, a requirement to preserve history might actually result in loss.

Type 2 SCD: This type creates a new record if the dimension changes and thus maintains a history of the values. This type will consume a lot of data if the dimensions keep changing very often.

Type 3 SCD: In this type, an extra column is appended to the dimension table, storing the previous value of the attribute that was changed. Thus, it can keep a history of values without data bloat.

The type of SCD you choose will depend on your specific needs and the requirements of your data warehouse.