This rule reorders group and window so it will switch from using `ReadGroup` to using `ReadWindowAggregate` when the intent is to aggregate a grouped window. It will then add a group node that groups by the given columns and the start and stop columns and then reperform the aggregate. This is more performant than performing the group first. |
||
---|---|---|
.. | ||
experimental | ||
influxdata/influxdb | ||
testing | ||
packages.go |