Take the 2-minute tour ×
Geographic Information Systems Stack Exchange is a question and answer site for cartographers, geographers and GIS professionals. It's 100% free, no registration required.

I was just asked in a interview what a dangling nodes was. I had no idea and asked a former co-worker.

I was informed that a significant amount of time was spent fixing these. I was confused how this could become a problem and came up with few quick solutions.

  1. Go through the system and identify all endpoints.
  2. Reduce the list of endpoints to endpoints that are terminals. (optional just less work)
  3. Eliminate all endpoints that aren't within a range of another object (within the same feature type)
  4. Create a report
  5. Have a system to mark the entry as resolved, if it is a valid entry (so it is not reproduced in the report). For example two roads that are really close but a 1 meter ditch separates them)
  6. Repeat regularly.

Solution 2:

  1. Check for system percolation
  2. Mark all entries that do not percolate within a small area
  3. Steps 4 through 6

This would produce a larger amount of false positives, but it's running time is in order (N log N)

We were just going through the map block by block looking for them. I must be missing something since I can think this problem can be solved relatively simply. I realize that the initial round would be quite a task, but with a system in place maintenance and monitoring would be minimal.

What am I missing that makes dangling nodes a large issue?

share|improve this question
1  
I've never heard of danglers. Are you sure they didn't mean dangling nodes ? The percolation link points to discussion about rasters. Usually dangles are dealt with in vector format. –  Kirk Kuykendall Jan 24 '13 at 22:19
    
Yes they did. They called them danglers. They also messed me up with asking me what an alias table was and not a table alias. Thanks for the correction. The percolation only shows a raster example. The theory is basically about values in a array which for some reason I now see analogous to certain trees and loop structures. The idea would be running the algorithm using a road network subsection as the entry and exit nodes as being the same node and check for a loop. I know it is missing a lot of checks, but I was hoping for a simpler example –  user13395 Jan 24 '13 at 22:35
    
Arcs will always have a certain number of dangling nodes; that is simply the nature of arcs. However you could have a 3 way intersection where one road does not form the 'T' and the node is considered a dangling node. Polygons can have dangles which can be selected and deleted. It seemed like more of an issue in the old ARCINFO command line and coverage world while editing polygons. There is a topology rule for the 'must not have dangles' now. The resulting list could be extensive but they are easy to identify with topology and correct. –  danagerous Jan 25 '13 at 5:18
    
Sorry. I'm assuming ESRI software. –  danagerous Jan 25 '13 at 5:19
add comment

1 Answer

up vote 2 down vote accepted

Without wanting to sound harsh, I am surprised that a GIS professional has not heard of dangling nodes as this was GIS 101 for me. I presume you also know about 'undershoots' and 'overshoots', 'bow-ties or knots', 'switchbacks' and 'spaghetti digitizing'.

Dangles are a big issue because they are bad topology and bad topology means your data won't perform as expected and the results of analysis may be wrong. For instance your road network will be broken because it doesn't realise a link exists between to road segments. Unclosed polygons are not polygons but merely lines. Bad topology also suggests inaccurate digitizing and therefore spatial error.

Rather than trying to re-invent the wheel, I'd use existing tools. Using appropriate snapping tolerances can help reduce the problem when digitizing but won't eliminate it. @dangerous mentioned Arc but GRASS, and therefore QGIS, also has tools for handling dangling nodes. Both Arc and GRASS also have tools for cleaning and re-building topology. These happen to be the GIS packages with which I'm most familiar but I would be surprised if the likes of MapInfo, Idrisi etc. don't have something similar as well.

Finally (from your comment), the term 'alias table' is sometimes used to refer to a table of aliases for geographic features and so would be different to a 'table alias'.

share|improve this answer
    
I am trying to become a GIS professional. I am a computer science professional trying to work my way backwards therefore constantly I am trying to map back basic terms. Thank you kindly for an answer filled with many of them. If you have a link/ISBN to this GIS 101 I would be most interested. All the beginner books I have picked up rattle on about how GIS is useful and other meta things and a long discussion coordinate systems with no math involved. Just to make you rage a little, I have an acquittance that that 0/3 GIS graduates for an interviews couldn't tell him what topology was. –  user13395 Jan 25 '13 at 12:48
    
Coming from a Computer Science background you will already know a lot but the language is a bit different as I have learnt working alongside programmers (for instance in GIS people often refer to Theissen Polygons, which are the same thing as Voronoi polygons). So it will often be a matter of translation from CompSci to GIS. My early training was on an ESRI course using the old version of ArcInfo when the Coverage data format was still in use, so I can't give you a link. However, try reading "GIS: A Computing Perspective" by W.F. Worboys. The book is a bit old, but still good. –  MappaGnosis Jan 25 '13 at 14:13
    
Thank you very much. I have never heard of either of those polygons. Purely relying on my math background (thank god for that) Comment worthy part: ISBN: 0415283752. –  user13395 Jan 25 '13 at 17:52
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.