Knowledge Drop

The PDT Build Process

  • 5 April 2021
  • 0 replies
  • 365 views

Userlevel 2

Last Tested: Feb 13, 2020

 

Whether a PDT is built by the query thread or by the regenerator, the build process is the same for all types of persistence.

Versions prior to 7.0:
The regenerator/query thread first builds a table with a name that begins with LC, short for Looker Create.

Once the LC table has been created successfully, if this is the first time the table is being built on the scratch schema, the regenerator/query thread will rename the table to a name that starts with LR, short for Looker Read. This LR table is the table that Looker will query from the UI.

If this is not the first time the table is being built on the scratch schema, the regenerator/query thread will first drop the previous version of the LR table and then rename the LC table to LR.

Versions 7.0+:
An LC table is no longer created prior to the LR table.

The regenerator/query thread builds one version of the PDT with a name that begins with LR. There is no longer a drop/rename process. Every new build of a PDT has a unique table name.

For all versions, if a regenerator rebuild fails, Looker will continue to use a previous version of the PDT until the regenerator is successfully able to rebuild the PDT.
Note: This does not apply to PDTs whose underlying definition has changed (i.e. a change to the sql).

More on PDT naming convention here.

 

This content is subject to limited support.                

 

 


0 replies

Be the first to reply!

Reply