I have 2 similar datasets which have some overlap. Long story short, I'm trying to combine data to and get it to match so that I can then build a Network layer from it. One of my problems though is that some of the overlapping data doesn't match exactly. An example below shows where both datasets contain the same road, but along slightly different alignments. I can't use any of the existing Select by Location options (I don't think) as I don't want just those that intersect or are identical, but those which roughly match the course of the selecting feature. Is there any way of doing this? Even if it's an extension, addin or python solution?
Solved! Go to Solution.
You tried a buffer around the "good" line of some finite difference or select by attributes within-a-distance-of ?
It won't be perfect but it looks like your data doesn't have a finite offset along the lengths or your examples and/or directions
You tried a buffer around the "good" line of some finite difference or select by attributes within-a-distance-of ?
It won't be perfect but it looks like your data doesn't have a finite offset along the lengths or your examples and/or directions
I haven't tried either of those - but I definitely will give them a crack. I'm not expecting perfect results and that manual cleanup work will be required, but considering I'm working with a good portion of Western Australias roads, any reduction in data cleaning will save a good amount of time!
Pay attention to any apparent consistency in the offsets, small differences (small being relative) could be due to a difference in coordinate system datum differences during the coordinate system definition process (eg in north america, "offsets" due to datum definitions for UTM and MTM projections can account for 1-100m offsets depending on the datum used for the same projection... you have them as well in Australia I assume)
I had the same problem lately.
What I did is to translate (very simple python) the lines to centroids and then running Near tool on the points.
If the lines are really similar (line length are similar) it is a quick and simple way to compare lines that are not exactly the same but almost. It also gets the case when one line have only two points and the other have many vertices that create straight line.
Have fun