I build road networks. Sometimes the data comes from different sources. One of the most time-consuming processes is to find lines that apparently touch each other but in reality do not, because they are not well drawn. They do not touch by very small amounts.
I guess sometimes this error happens when the person drawing the line forgets to snap properly. I've also seen people (me included) moving lines inadvertently by small amounts when clicking on them in an editing session.
I know I can select all dangles in a topology, but many of them (actually, most of them) are "valid" dangles, that is, places where a road actually ends. So, when you check the topology you end with a lot of "false positives".
In short, I want to find the "false" dangles, that is, lines whose end is at a small distance from another line because of drawing errors.
As you can imagine, when I say "small distance" it means one thing, and when other person says the same he can mean a different thing. I can be talking of meters, millimeters or kilometers, it depends on the application you're working in.
I just created a topology, took the dangles that result from the topology, converted them to a layer, created a buffer for the line feature and found which dangles fall within the buffer. As usual. It is a time consuming process and you are never sure you've found all errors.
This process took me the best part of today, given the complexity of the network I'm working in and the amount of editing it has "suffered" by four different contractors (and the topology editing I had to make because of other errors).
After checking for lines at less than 5 meters, I just found a site where the lines are 10 meters apart, God knows why. Then, I threw my hands up in the air and started this thread.
Simple: has anyone ever achieved such a revision of a network in less than half a day?
Why is that I cannot find a tool or topology rule that allows me to do such a process? I mean: to find lines whose non-connected ends are at an arbitrary distance from another line. Please, enlighten me.