2

I have run into this before and simply marked it as an exception and moved on, but it puzzles and bothers me a bit so if someone can provide a reasonable explanation, that would help me. Surprisingly, I couldn't find a post related to it here or in Esri's online documentation.

When I create a topology and add the rule "Must Not Have Gaps", even with this 2-feature polygon feature class, there seems to always be one error that encompasses the entire feature class.

Can this error be effectively fixed?

When I have it create a feature (a common fix for topology issues), it creates a single polygon that includes the area of both features but not the boundary between them and the error does not appear to include any part of the boundary between the features as you can see in the screenshot.

I remember in the days of "Coverages" and ArcInfo Workstation that there was this concept of a "World Polygon" that represented the universe outside of your extent (and you were NOT to delete it). The only "reason" (for lack of a better word) for what I am seeing is that this is a manifestation or result of that same concept. Maybe it should be marked as an exception and I should just move on as I've done in the past. Comments suggesting things to try (especially if this isn't exactly the "non-error" that I've described), are welcome. I'm using ArcMap 10.6 for this.

enter image description here

9
  • 4
    Mark the error as an exception and move on. Your polygons are not filling the entire coordinate extent of the feature class and GIS has no way of knowing that this is the outermost extent of your particular area of interest without you telling it that you have reached your limit. If this was a jurisdictional boundary, you might get data from an adjoining jurisdiction to do a joint project or study and once you merged their polygons with yours the topology should identify your normal outermost boundary as having gaps with those new polygons. Then your normal exception becomes an error. – Richard Fairhurst Jul 11 '19 at 16:43
  • So your title is not correct, since only a polygon or set of polygons that encompass the entire coordinate extent of the feature class can genuinely be called "the entire feature class". – Richard Fairhurst Jul 11 '19 at 16:52
  • Ok so when would this not happen? It seems to me that being square or rectangular wouldn't satisfy the rule entirely either, so I'm still missing something. I will change the title slightly from "entire feature class" to "extent of all existing features". – jbchurchill Jul 11 '19 at 19:00
  • 3
    There is no real edge that can ever be excluded universally, so topological errors will always happen when ever you define an outer edge. The outermost edge of every 2D global coordinate system is not real. Its artificially imposed by the author as a convenience that serves the intended design of that particular representation of a 3D sphere. All 2D coordinate systems can be exchanged for another with a different artificial edge, since every edge you can draw on the face of the globe has a topological relationship to something on both sides of that edge everywhere in the real world. – Richard Fairhurst Jul 12 '19 at 2:02
  • 1
    I will repost my comments as an answer with a picture that illustrates the things I described in my last two comments. – Richard Fairhurst Jul 13 '19 at 2:09
4

The answer to the question "Can this be effectively fixed?" is essentially no. The picture below illustrates why it is a good thing that the outer pink boundary is created by the topology and displayed and why only the user can determine if the exterior boundary shown contains an error or is as correct as it will ever be and is a topological exception.

enter image description here

The pink line shown is all part of the exterior edge of the two polygons. The portion of the pink outline that seems to extend between the two polygons like a single straight line at this zoom factor might in fact be a gap that is perhaps 3 meters wide and 30 km long. This gap could be a topological error in the boundary created by merging two feature classes together that were independently created without topological control. However, this gap could represent the actual exterior boundary of these two polygons if the gap excluded a utility easement that belonged to another jurisdiction or was part of a gerrymandered political boundary that politicians have created.

If the topology automatically treated this topological boundary as an exception simply because it was outside of the interior extent of all existing features, the user might fail to zoom in to the actual gap that exists based on the incorrect assumption that the lack of an error means that the entire straight edge between the two polygons was a shared boundary between them. However, because this boundary is always shown unless the user explicitly marks it as a topological exception, it draws the user's attention to this gap.

If the user determines this gap formed by the exterior boundary of the features is an error, the topology provides tools that can assist the user in closing that gap or the user can fix the gap using other methods, such as cogo. Only after the user has modified the boundary to their satisfaction and validated it by rerunning the topology or determined that this outline is already as correct as it will ever be without closing the gap should this topological error be marked as an exception. Only the user can determine whether or not this boundary actually violates or is an acceptable exception to the "Must Not Have Gaps" topology rule.

Topologies are only effective if they draw a user's attention to all potential errors that violate the strict enforcement of the rules that the user chose and assist them in fixing those errors or permit them to mark them as an exception when the strict interpretation of the rule does not fit the user's needs.

1
  • Thanks … I think this is a perfect explanation. I’m going to look through this a little bit later when I have some time, to see if I have any further questions, but I think I understand all your comments so far. – jbchurchill Jul 14 '19 at 12:28

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.