In Redshift we often need to split a huge event table into different smaller table. The objectives are to increase performance and facilitate end users analytics queries. Let’s also assume each event row belongs to an event type and that we have time series data.
What are the best practices regarding such an endeavor and the common pitfall to avoid?