How does storing data by columns improve query performance?

[Andamantourpackage]

Columnar storage is one of the reasons that big data can be handled well with analytics systems. The data is not grouped by the entire row, but by the individual column; this means that a query can only read the data that it requires. This can significantly decrease the amount of data that is triggered for scanning, especially if the tables have a lot of columns. It is also efficient to have similar values grouped together in columns, and this can further enhance performance. It is particularly valuable for the aggregations and large scans in analytical workloads. Again, the proper option will depend on query patterns, data size and the primary purpose (mostly analytics or transactions).