1

I have the output of the route layer. And I want to determine the number of right and left turns for each route. For that, I used the COPY TRAVERSED ROUTE FEATURES tool from Network Analyst. However, this gave an empty output table for the turns. Are there any prerequisites for the network dataset to use the COPY TRAVERSED SOURCE FEATURE tool? Or is this possible by using Global Turns in the analysis by any way?

I am looking for the result similar to below (from a study of Dalumpines and Scott).

enter image description here

  • I cannot say if this is the method they used to generate that table. I can answer the main question of why the tool didn't work for you. I think this question is sufficiently different from the previous one (where you asked for any method), although you could have edited this in to it as something else you had tried. I would leave both questions - if this method works for you, it could be posted as an answer to the other one, but the question/answer here is more about why the tool produced an empty output. – Chris W May 31 '15 at 20:55
2

When considering turns as features in a network dataset, they must be modeled first. That means you have to create an actual turns feature class that has its own geometry to be traversed. When turns are not specifically modeled (ie, the Global Turns properties you mention) they aren't features that are traversed, just analysis variable properties, which is why your table is blank.

Usually turns are only modeled in special circumstances, for instance when they need to be restricted at a particular intersection. However it is possible to model every turn in the network. I've never attempted to do this on my own, and I'm not sure if there are any automated tools to generate the turn features for an entire network at once. Before committing to creating a full turns feature class, you might just create a few sample ones that would fall on your routes and then use that to test if it produces the desired attributes or something useful to work with.

I cannot say if those attributes in your example image were generated using this tool/method. Based on the Copy Traversed Source Features help files, and specifically the output documentation, it does not appear that this tool will generate the information you want. I also found this GeoNet forum post with the same goal as you, and my interpretation of the information there is the tool doesn't produce that information. (The suggested solution there is also to use Directions, and explains briefly how it comes up with left/right.)

  • I tried to find if this will give left and right turns for only one route by manually selecting the two graphic location. But, although the "turn" table have outputs its not what I needed. Now, I am trying to implement the turn restriction (10 sec for right turn and 30 seconds for left turn). I am trying to find the number of turns on the basis of time value. However, I havent reached the way to do this. Thank you for the suggestion. – Droko May 31 '15 at 21:37
  • @Ranjit - answer updated with new info. I don't think the restriction thing would work. For one, the added restrictions to those turns might alter the route. Even if it didn't affect the route, you'd need numbers that could never sum to a value of the other type. The GeoNet forum makes the same suggestion I originally did - use route directions, and parse that output xml file. – Chris W Jun 1 '15 at 4:14
  • Yes. That is correct. The use of restriction yielded the time-value which was impossible to use for detecting the type and number of turns. I went through the geonet forum on the similar kind of problem and the solution was to parse out the XML file, as you did initially at my previous post. Does it require heavy coding if I want to get the number of Left and Right from the XML file for each of the many trips? – Droko Jun 1 '15 at 17:34

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.