2.1.0 cost/reverse_cost not set correctly for oneway streets

[woodbri]

using This is osm2pgrouting Version 2.1, I just noticed that cost and reverse_cost values are NOT set to 1000000 for the wrong way on one_way streets. The same for cost_s and reverse_cost_s. In fact it is even worse than that because the reverse_cost for one_way=1 are negative values which might be ok if the algorithm eliminates these edges, but when one_way in (0,2) both values are positive.

417;102;0.000577722511245135;497.286624538597;"";16329;5454;-71.3205203;42.6074975;-71.32002;42.6077864;0.000577722511245135;-0.000577722511245135;13.8135173482944;-13.8135173482944;"";1;130;130;8888815;64835130;64910422;1;""
436;112;0.00198361736734031;165.789685892489;"Saratoga Street";3266;11195;-71.3019915;42.6186827;-71.3000534;42.6191052;0.00198361736734031;-0.00198361736734031;12.7530527609607;-12.7530527609607;"";1;50;50;8888881;64931814;64894445;2.5;""
461;102;0.000561720330772339;57.4187938952264;"";11650;13358;-71.3281444;42.6132963;-71.3284706;42.6137536;0.000561720330772339;-0.000561720330772339;2.39244974563443;-2.39244974563443;"";1;88;88;8888835;64929542;64926257;1;""
474;108;0.000458349920919023;38.2496018886591;"Merrimack Street";11803;15493;-71.307384;42.645424;-71.3078351;42.6455052;0.000458349920919023;0.000458349920919023;1.52998407554636;1.52998407554636;"";2;90;90;8888850;64910689;1236794932;1.5;""
475;108;0.000507758259406249;41.8497802544357;"Merrimack Street";15493;13171;-71.3078351;42.6455052;-71.30834;42.645559;0.000507758259406249;0.000507758259406249;1.67399121017743;1.67399121017743;"";2;90;90;8888850;1236794932;64855170;1.5;""
476;108;0.000922436447669103;75.8136535115301;"Merrimack Street";13171;11790;-71.30834;42.645559;-71.30926;42.645626;0.000922436447669103;0.000922436447669103;3.03254614046121;3.03254614046121;"";2;90;90;8888850;64855170;64820931;1.5;""

[cvvergara]

If you intend to use the values an algorithm that don't correctly creates the directed/undirected graph based on the fct that negative values are not inserted, then you can modify the generated negative values (the edge does not exist) to be highly costly.

[woodbri]

There is clearly a bug here because the software does not behave the same way for one_way=1 vs. one_way=2.

• for one_way=1 reverse_cost is negative
• for one_way=2 I would expect cost to be negative but both are positive like one_way=0
• for one_way=0 both are positive as I would expect

I presume one_way =0, both ways ok, =1 from-to is ok, =2 to-from is ok. But I could not find where this is documented.

[cvvergara]

Probably the documentation made by GSoc Student is incomplete

     " one_way int, " 
 -1 reversed (1 way but geometry is reversed)
 0   unknown,
 1   yes(normal direction),
 2   not one way its 2 way,
 3 - reversible (one way street but direction changes on time)

0 & 2 & 3 put positive values on all cost /reverse_cost -1 has negative value on cost column 1 has negative value on reverse_cost column