My polygon featureclass (in a file geodb) has 1531 polygons. It was originally created from a raster so the linework is blocky. I have already eliminated all tiny polygons, dissolved, repair geometry.
I am trying to run Simplify Polygon as follows: CA.SimplifyPolygon (inStands, simpStands1, "POINT_REMOVE", 2, 2000, "RESOLVE_ERRORS")
I have tried it with different parameters and it still hangs. If I choose "NO_CHECK" instead of "RESOLVE_ERRORS", it runs very fast, but just creates a huge mess - not only slivers, but some of the original linework is still there doubled up with the new linework - it's like it didn't really finish. "RESOLVE_ERRORS" does a beautiful job on a subarea, but hangs on the whole file. It hangs at the stage "Assembling Features". It has created some temporary line feature files but the output featureclass is still empty. It hangs forever and I have no choice but to crash it.
I had the same problem with another similar file, but eventually I got it to run at the above settings. I can't go any higher with the point remove tolerance or the lines get too generalized.
I really need to simplify these polygons. Splitting it up is a last resort as the edgematching after simplify and smooth will be ugly.
1. Run the tool without the 2000 (min. area) and see if it runs through. You can use Eliminate Polygon Part tool , assuming you have ArcGIS 10 arcinfo license, to remove small holes or polygons before or after Simplify Polygon. I would do that before simplification to make the input to Simplify Polygon less complex.
2. Run the tool with a tolerance smaller enough to not to generate a mess (line crossing, etc.) , for example 0.2 or 0.5, in a first round. If the tolerance is safe you can try the NO_CHECK option. If it runs through, run a second time with the desired parameters and option.
What is your data projection and units? Hope it is in meters or feet, not decimal degrees.
Dan, your suggestion #2 almost worked - i.e. after first running Simplify Polygon with a very low tolerance and NO_CHECK, and then running it again with desired parameters, the tool did get further in the process (was able to process more features) before hanging. But it still hung.
The problem with No_check is that, even at the very low tolerance, it creates slivers that are actually gaps in the data - empty holes. I can't remove them with Eliminate because I can't select them. I'm not sure how to get rid of them at all.
I should have suggested running the tool with a very small tolerance using the RESOLVE_ERRORS option so that it won't create gaps. The main reason for trying this is to force the features being modified and see if that makes the second run go through. You can give it a try one more time. Try this a couple of times. If it always hungs and if an incomplete output is left in the gdb when you kill the process, have a look of the OBJECTID of the incomplete table. Does it hung after the same feature OID? Select the feature of the next OBJECTID in the original input to process only that feature using the desired parameters. Does the tool work for this feature? If the tool doesn't work for that feature, try to select all other features to process.
If nothing works and if the tool hungs at different features (inconsistent), I suggest that you try to reduce the data to a sample area that reproduces the case and submit a bug to esri Support.