Dynamic Timeframe

Standard dynamic timeframe solution described here: https://help.looker.com/hc/en-us/articles/360001288468-Dynamic-Timeframes-for-Dimension-Groups has its drawbacks. Because the resulting dimension is string, it means it won't get an automatic formatting from Looker, i.e. January instead of 2022-01-01

However, there is a way that allows us to use dynamic timeframe and granularity, and also adds missing features:

1. You can filter using timeframe field rather than the field you're joining to

2. Dimension is returned as date, which means its recognised as time series

3. Application of the code can be simplified so that you don't need to copy it everywhere

Steps

High-level steps are as follows:

1. Expose your calendar table as view calendar.view.lkml

2. Create a refinement of calendar view called parameters.view.lkml

3. Join the calendar to your view

4. Change view_label of the calendar join to your main view's name

5. Enjoy

Things to remember:

1. In order to be able to use Timeframe with other values like Day of Week or Week of Year, the code would have to be altered.