I would also add about limitation and how to overcome it:
- copy of json vs csv (Non Strict/Strict schema)
- limitation of udf (no input possible)
- listagg on more than varchar(max) using ...
- some redshift sql tricks like (equivalent to generate_series) https://github.com/eyaltrabelsi/my-notebooks/blob/master/tot...
- Emphasize lack of recursive functions& triggers and more
and how one would fix it using code other tools
ERROR handling:
- SSL
- load errors
and some more advanced stuff like:
- dynamic schemas since data+quries alive how would one generate schemas
- wlm and the new features
- auto scaling the cluster when/how few slices with big storage vs many slices (Network IO vs parallelize)