I would like to sum the amount of customers like on the picture added but for example if the date is 2009-03-09 the the customers that have been visiting between 2009-03-08 19:00 and 2009-03-09 19:00 are summarized. Can I achieve it with set-analysis?
As far as I know, you can't... If my understanding is correct, the Set Analysis expression is being evaluated "outside" of the context of the current cell in your chart, therefore we can't use the current dimension value (e.g. Date) in the set analysis expression.
I'd be very happy to see a solution to this problem, because problems of this kind have been bugging me for a while...
As far as I understand your problem, you can either use an IF statement (which is slow, if you have a lot of data), or solve this problem back in the script and associate the relevant timestamps with the desired Dates.
I agree with Oleg that you would need to do it in the script. Come up with a routine that will group those time periods into the appropriate groupings/buckets. Probably can use intervalmatch.
Could you provide us with some sample data to test it out?
