- or -
You can create incremental PDTs in your project if your dialect supports them. An incremental PDT is a persistent derived table (PDT) that Looker builds by appending fresh data to the table, instead of rebuilding the table in its entirety. See the Incremental PDTs documentation page for more information.
increment_key parameter specifies the time period for which fresh data should be queried and appended to the aggregate table. The
increment_offset parameter is an optional parameter you can use if you want to rebuild the table for previous time periods at the same time that fresh data is appended to the table. The
increment_offset parameter defines the number of previous time periods that will be rebuilt when appending data to an incremental PDT.
increment_offset parameter is useful in the case of late-arriving data, where previous time periods may have new data that wasn’t included when the corresponding increment was originally built and appended to the PDT.
increment_offset value is
0, which means that only the new data from the current increment is appended to the table. If you set the
1, late-arriving data from the previous increment will be added to the table in addition to the new data from the current time increment.
See the Incremental PDTs documentation page for some example scenarios that illustrate how incremental PDTs work and that show the interaction of
increment_offset, and persistence strategy.
See the Supported database dialects for incremental PDTs section on this page for the list of dialects that support incremental PDTs.
For example, this PDT is rebuilt in increments of one day (
increment_key: event_day), going back three days (
Supported database dialects for incremental PDTs
For Looker to support incremental PDTs in your Looker project, your database dialect must support Data Definition Language (DDL) commands that enable deleting and inserting rows.
The following table shows which dialects support incremental PDTs in Looker 21.16: