Which data model is preferred for Vertica to store key value pairs? Say, we read data from 1000 sensors (all at the same time) every 30 seconds. So after a year, we have one million data points of vectors of size 1000. So we have full dataset, i.e. dataset is not a sparse matrix.
Option 1: Store as key-value pairs (key=sensor_id, value=sensor_value)
Option 2: Have a table with 1000 columns, i.e. one column per sensor.
In general, what's the cost of having a table with too many columns?
Thanks!
