Column Storage and Compression in Data Warehouse
Unlike standard row-based relational databases, the Data Warehouse stores data using a columnar storage mechanism:
Columnar storage is particularly suitable for improving disk performance when retrieving complex analytical queries or running in-database business transformations. Queries can be answered by accessing only the columns required by the query, which fits well with data warehousing and other read-intensive use cases.
Moreover, the data in columns are compressed using various encoding and compression mechanisms, which further improves the disk I/O. For example, run-length encoding can be applied to the “symbol” column.