The Informed Company

Chapter NineteenRow‐ versus Column‐Oriented Database

We wanted to spend a bit of time talking about why columnar databases are more performant for analytics, since this will be at the center of your data stack. There are two ways to organize relational databases:

• Row‐oriented
• Column‐oriented (also known as columnar or C‐store)

Row‐oriented databases are databases that organize data by record, keeping all of the data associated with a record next to each other in memory. Row‐oriented databases are the traditional way of organizing data and still provide some key benefits for storing data quickly. They are optimized for reading and writing rows efficiently.

Common row‐oriented databases:

• Postgres
• MySQL

Column‐oriented databases are databases that organize data by field, keeping all of the data associated with a field next to each other in memory. Columnar databases have grown in popularity and provide performance advantages to querying data. They are optimized for reading and computing on columns efficiently.

Common column‐oriented databases:

• Redshift
• BigQuery
• Snowflake

Row‐Oriented Databases

Traditional database management systems were created to store data. They are optimized to read and write a single row of data, which leads to a series of design choices, including having a row store architecture.

In a row store, or row‐oriented database, the data is stored row by row, such that the first column of a row will be next to the last column of the previous row.

For instance, ...