Knowledge Drop
Last Tested: Mar 27, 2020
If a PDT is defined in production LookML and the PDT is persisted using the datagroup_trigger or sql_trigger_value parameter, the regenerator typically handles the creation of the PDT on the scratch schema.
The regenerator is the process in Looker that manages the building of PDTs on the scratch schema. At a high level, the regenerator process works by checking the current value returned by the sql_trigger_value/sql_trigger of a PDT and comparing that value to the previous value returned for that PDT. If there has been a change in this value, the regenerator will rebuild the PDT on the scratch schema. More on the regenerator process here.
There is one exception to trigger-based PDTs in production being managed by the regenerator -
If the PDT is queried via an explore/look/dashboard and it has not yet been built (or rebuilt if the previous build is expired) by the regenerator, the explore query thread will handle the creation of this table.
LinkedIn
Twitter