Skip to main content

Hypertable Types

Scheduled Hypertables​

Scheduled HyperTables are like “materialised views” which are kept updated at regular intervals decided by the user. Scheduled HyperTables are like materialised views. The backing data is stored internally in Z-Lake (which is based on Apache Iceberg) and is available instantaneously when SQL queries are run on it.

tip
  1. Should be the first choice unless there is a need for the data to be “live”
  2. Works well with both types of data sources - databases and SaaS sources
  3. Very fast results when compared to live mode

Live Hypertables​

Live HyperTables are like “views” which provide data directly from the underlying data sources as and when required. The data is not stored anywhere and is queried in-place and transferred to the SQL client directly.

tip
  1. Should be avoided unless there is a need for data to be “live”
  2. Works well with databases based data sources (MySQL, PostgreSQL, Redshift, Snowflake, Big query etc)
  3. Do not use with SaaS sources unless absolutely necessary
    1. Since rows are coming in the SaaS application, this will be slower
    2. If a huge amount of data has to be consumed, API rate limitation / throttling from SaaS provider might affect performance