This is work in progress as of May 2021, not all examples are complete and things are changing. I organize my thoughts into five easy pieces below.
Slides from a talk on this topic at the Cleveland R Users Group:
The easy pieces:
- Large out of core data
- TPCH join-aggregate example from DuckDB
- Last item per group
- Genomic overlap joins
- "As of" joins
A SQL rant born out of frustration while compiling these notes appears here: