I have 35m rows of log data from Expedia, which has 3 date columns. As part of my feature engineering I want to break this in to Year, Month, Day, Hour, Day of Week and Quarter, then feed this into a ML model.
Month, Day, Hour, Day of Week are commonly extracted features (and already provided), however in this case I suspect Quarter will be more important as the usage (hotel booking) has seasonality that follows the Quarters. .
Clearly I can do this using apply and a calculation from the month, however I thought Quarter would be a nice addition to split_datetime and is already standard in panda.