I wish to insert data into a table only when the value of the inserted data has changed. Thus, in a time series, if the value of the data at time, t-1, is 206 then if the data to be inserted at time t is 206, then it is skipped (not entered). If the value of the data at time t+1 is 206, it is skipped also; until the value changes, so if the value at t+1 was 205, then that would be inserted, and if at time t+2 the data is 206, it would be inserted too.
What is the best way to do it without increasing overheads?