Open
Description
Processing optimizations:
- Have the list of events sorted by timestamp, then we could binary search the events intersecting. Doing so would significantly improve performance of transforms like filter_period_intersect and the union_event_split in aw-transform: Add union_events_split #179 .
Memory optimizations:
- Wrap Vec<Event> or whole queries in an allocation arena
- Reduce the amount of clones in our transforms (already been improved quite a bit the past few months)
- Add proper lifetimes on variables in query2 to avoid a clone on each variable assignment and reference to it (See Add explicit cloning to query2 #119)
Something else?