Select to view content in your preferred language

Export Web Map and legend color oddness - ArcGIS for Server 10.1

502
1
02-19-2013 10:18 AM
FrankMcLean
Occasional Contributor
Hi,

I have a geoprocessing service which exports its results as a map service.  I use this to produce a printable map via the Export Web Map utility.  The issue I'm having is the seemingly odd designation of colors for the output layer.  I've symbolized the output with the exact same layer file I use for the input (output is basically a subset of input) and the colors are completely different, though consistently (as far as I can tell) completely different, i.e. the same odd palate is used each time.

Does anyone have any clues?

Thanks,

Frank
Tags (2)
0 Kudos
1 Reply
FrankMcLean
Occasional Contributor
I'll provide my own 'solution' here. 🙂  We're talking ArcGIS 10.1, no service pack, if it makes a difference.

The problem is that publishing only takes 'half' of the symbology for a more complex case.  Example:

I have two GP output feature sets.  One simply symbolizes all shapes with a border color, irrespective of feature contents.  The other symbolizes based on a type field within each feature.  Both symbologies are set directly on the ModelBuilder output parameters.  When the model is run in ArcMap, the results are correctly symbolized on screen for both feature sets.

It's publishing the result that is the problem.  If I publish the result without changes, the first, simple map service symbolizes correctly.  The second, complex case, however, takes only 'half' of the symbology.  The features are correctly discriminated in terms of the type field, so the break is correct and the features are identifiable by type, but the colors are not carried over.  Rather, a default set of colors seems to be applied.  Meaning that I can see something is a 'plant', but it's bright red, instead of green.  This is particularly problematic if the original layer that was subset to produce the output already uses bright red for another type.

The solution is to re-apply the layer files to the outputs on the map before publishing the result.  I know this is what the docs say to do, but I don't understand why the 'half-cooked' situation is happening.  Well, I suspect a bug, but there you go.  It would be wonderful if a model could fully encapsulate the treatment of its outputs.

Thanks,

Frank
0 Kudos