Data analytics teams, plenty of times, would have to do long-term trend analysis to study patterns over time. Some of the common analyses are WoW (week over week), MoM (month over month), and YoY (year over year). This would usually require data to be stored across multiple years. 

However, this takes up a lot of storage and querying across years worth of partitions is inefficient and expensive. On top of this, if we have to do user attribute cuts, it will be more cumbersome. To overcome this issue, we can implement an efficient solution using datelists.

Leave a Reply

Your email address will not be published. Required fields are marked *