I am new to knime,
help please, which nodes can I use to solve the problem of the previous image.
the column id, datetime, price are input values, sum_price is output data.
I must add the prices of a previous day of each row and assign in the new column (sum_price). “?” is assigned because the date 01/01/2000 has no previous day.
could you provide an example and say again what should happen with the sum_price. What does get aggregated in which way? If you want to do things with previous or following rows there are several principal options:
What I have figured out from your print screen is following. First appearance of day (for example row with id 5) you have to sum everything from previous day (81), every other appearance of same day (rows with id 6 and 7) you sum everything prior in that day in addition with sum of previous day. So for row with id 6 you have 81+34 (=115) and for next you got 81+34+12 (=127).
Basically you want to get sum of each day and cumulative sum in each day. Lag both and then sum it. Explore workflow, play a bit with it using different configurations and if any questions feel free to ask.
I’m confused. It seems that you are interested to get the sum of rows within the last 24 hours. Is that correct?
If it is so, then it contradicts the example that you have provided: row 5 should include only row 4 in the sum (thus outcome should be 23 instead of 81), as row 3 is more than 24 hours in the past (same as row 11 vs 14). It seems to me, that the previous answer by @ipazin fully reproduces your original request (other then outputting missing values for ALL rows of the first day)
it is correct, sum of rows within te last 24 hours.
It is correct row 5 of the column sum_price should be 23 (typing error), the example @ipazin is fine only that sum by dates and does not consider the time of 24 hours.