Hi Tom, Good that you checked the problem formulation - forces me to improve the structuring of the problem! It is not a simple "travelling salesman" problem, the fastest route between two points, but rather the closest match, to a target stop, along a PT route. Thus, the first table of Message #9, which consists of only one card transaction, contains, in column #8, possible (candidate) stops based on network _walk_ distances to stops surrounding to stop in the 7th column (#7400003), which is the boarding stop of the _subsequent_ (next in time) PT trip (i e subsequent card transaction in the ticketing data). Note that all the rows in ths sample of the first table have the same values of column1-6 and thus belong to the same card transaction, with the boarding of the subsequent trip appended to it. (The boarding stop of the transaction itself is thus 7401547 - column #3). What about the second table then? Well, it contains the trip itinerary of line route 817, in this case identified by the line route ID of the first column in table #1, and looked up in the trip itinerary database based on column #2 (table #2 of message #9). Now, the script should loop through the stops of column #8 of table #1 and compare them to the stops of column #4 in Table #2 and indicate the stop (in column #8 in Table #1) with the smallest value of either column #9 OR column #10 in Table #1. This should result in a '1' in column #10 of Table #3, which is just a reformulation of Table #1, plus the "flag" column. As to what is missing in the code of the datasteps i provided in message #9, I am still lost in oblivion...
... View more