Compression

Time-series data can be compressed to reduce the amount of storage required, and increase the speed of some queries. This is a cornerstone feature of TimescaleDB. When new data is added to your database, it is in the form of uncompressed rows. TimescaleDB uses a built-in job scheduler to convert this data to the form of compressed columns. This occurs across chunks of TimescaleDB hypertables.

This section explains what native compression is, and goes through some of the benefits and limitations of compression. It also includes instructions for setting up compression, and using it in your environment. We strongly recommend that you understand how compression works before you start enabling it on your hypertables.

• Learn about compression to understand how it works before you begin using it.
• Manually compress specific chunks.
• Decompress chunks to manually decompress specific chunks.
• Backfill historical data to insert a batch of data into a compressed chunk.
• Modify schema to modify the table definition for a hypertable with compressed chunks.

warning

Compression alters data on your disk, so always back up before you start!