Edge Snapping

[LisaDygert]

Hello,

I am adjusting some polygons so that adjacent areas that share a boundary have coincident boundaries.  Do to some issues with other feature class dependencies, I cannot always add vertices to existing polygons to snap to.  I am trying to use the edge snap to resolve this, but if I zoom in really close it appears that the two areas are not actually snapped together.  Is this actually the problem or does it just appear that they are not snapped together?  Is there a setting I can use to make it actually snap or make it appear snapped if it is?  I know if I am able to snap to existing verices and I zoon in to the same scale it does appear to be snapped.  I am concerned as to whether it actually is snapping correctly with the edge snapping option.  I am using version 9.3. 


Thanks,
Lisa Dygert

[DonovanCameron]

You can use the integrate tool to edge-match polygons.

Or, place your data into a geodatabase and establish some topology rules and validate them. Learning how to manage topology rules and relationships inside a geodatabase can shave tonnes of time from almost any editing task.

Topology Rules Poster (pdf 1.9mb)

[MarsSjoden]

I like the "Integrate" tool for this purpose.  Lets me be a little more messy, less fussy on creation.

Not sure if there is a licensing req. to use it.

Integrate Tool

[MichaelStead]

It depends.

If one feature is more accurate than the second, integrate is not ideal as it modifies both features. If both features are equally precise/accurate it is quicker/easier.

If you do this with topology rules at least you have input over what geometry gets modified, and this might be important when you are matching to existing data that you don't want modified (exists in maps, databases, reports, etc)

A quick check of a shapefile to look for gaps is just to select and merge all polygon and see if there is more than one record in the sketch geometry vertices table. To look for holes you can create a square polygon of your data extent and erase it with your data then look for slivers that were not erased.

ET Geowizards also has a function called Eliminate that works good to clean things up a little less manually, it is somewhat better than integrate at letting you define what you want to allow to be modified.

[DonovanCameron]

Integrate is one of the few tools that will allow Priority ranks where features can be processed with a given hierarchy. A selection is also a good way to set up some ranking; only selected features will be modified on output.

There is an Eliminate tool in ArcGIS also, but restricted to the ArcINFO licensing level. Good ol' Geowizards (always inspect your results when using this one, closely...)

[MichaelStead]

I thought I remembered that even if you prioritize your snapping preference it still modifies both polygons somewhat. I remember trying to get some new TEM data incorporated into an existing layer and not wanting any changes in area from the old polygons and having a heck of a time. Some time in the past now, I might have been the ET Geo that was giving me the problems..... could easily be mistaken. Something to think about when doing these type of edits anyways...

Any specific Geowizards problems you could mention Donovan? I use it all the time as I only have an ArcView license.......so I can pan and zoom and use the id button 😉

[DonovanCameron]

Any specific Geowizards problems you could mention Donovan? I use it all the time as I only have an ArcView license.......so I can pan and zoom and use the id button 😉

I noticed some problems primarily when converting from lines to polygons or vis-a-vis.

I haven't used it since 9.1 so maybe it's been upgraded, but zoom in to see, for example: if converting a closed polyline to a polygon that the polygon matches completely with the polyline

Because sometimes I am faced with edge matching that requires only a single edge to match and converting to line and breaking it into manageable segments can make the job a little quicker (than re-assemble to polygon/feature class). Like replacing an edge with a more accurate one, sometimes done for coastlines.

[MichaelStead]

Thanks Don,

It does have some issues with both the build and the convert polylines to polygons for sure, but with a few checks in place I can get pretty close to 100%. I get basedata as linework and if I get the odd swamp classed as a marsh it isn't the end of the world. Most work I do that is critical would not be old school digitized as lines and built anyway.

Thanks for responding. I like to do everything myself, but that is impractical so i depend on my couple of hundred dollar license for Geowizards to take care of everyday stuff that I don't have the $$$ ESRI license to do.

[JohnSobetzer]

Lisa, I think you have two different issues.  One is the visual aspect.  I think ArcMap lets you zoom in past 1:1 but in my experience beyond that what you see is not what you get.  The other is the topological/snapping aspect.  If you are attempting to move a vertice of one polygon to snap to a line of an abutting polygon with the idea of having their edges be perfect coincident, there may be a problem because of the precision.  That vertice has to "snap" to some underlying grid and that point may not be "on" the line.  I think to have perfectly coincident shared boundaries in shapefiles and feature classes you have to have vertices in the same place. Having said that the difference is probably so small you wouldn't notice it, and you probably can't see it due to the visual limitations.  And even if you could, as one of my coworkers said, you can fix errors or you can make your lines thicker and/or change your scale.

If you create a topology and verify it I think a matching vertice will be added along the line of the abutting polygon.  Or you could clean your layer after editing.  Introducing a vertice along a line by either approach may cause it to "jump" in comparison to the backup dataset without that vertice, however. Or you could convert your polygons to lines and points, clean the lines, and from then on edit them and build polygons using the lines and points, the points providing the attributes.  Then a single line is used as the boundary for abutting polygons much like a coverage works.