Unforced change/error in symbology of feature services in web map & app

3520
7
11-30-2015 10:12 PM
HCCGISICT_
New Contributor III

Hello,

Our organization has a number of web maps feeding apps which are on our public maps gallery.

Recently, on this particular map:

https://hobartcc.maps.arcgis.com/home/item.html?id=f32a803c066a457aaafa73c6003be761

Two feature services have suffered symbology failures with no interaction on our behalf.  Content of feature classes has remained untouched since initial publishing in early November 2015.

  • Costal Hazard Inundation
  • Costal Erosion Hazard Bands

I have tried removing the services from the map, saving the map, deleting the feature services and feature definitions, republishing and adding the new services to the map with no change.  An attempt to add the republished feature services to a new map window yields a similar result. I have also tried running a topology check for errors with nothing jumping out as being an issue.

I am able to export the feature services to shapefile in arcgisonline, download and open in ArcGIS desktop with no symbology errors.

Any advice would be greatly appreciated.

Regards,

Roy Langman

City of Hobart

Tasmania, Australia.

0 Kudos
7 Replies
PaulTrevillion
Esri Contributor

Roy, your question caught my interest and so I took a closer look at your data. I noticed that some Coastal Hazard Inundation polygons are extremely small. For example, OBJECTID=115 is just a few square centimeters. Even though these sliver polygons are, presumably, unintended, they shouldn't normally be a big problem. However, I suspect that the projection from 28355 to Web Mercator (102100) at this scale is somehow screwing things up.

To quickly test this notion, I think I would make copies of the two suspect data sets and simply delete all polygons that are less than, say, 1 square meter. Point the layers to these edited copies and republish the services. Test the web map and if it solves the problem, you could then look at more sophisticated editing techniques to preserve the areas from the slivers in your original data.

0 Kudos
HCCGISICT_
New Contributor III

Hey Paul,

Your fast response and advice is fantastic! I was really scratching my head as to what may be causing the issue.

I will try your suggestion ASAP.

These suggestions are helpful and will hopefully correct the current issue.

Has there been a change to the projection system used in arcgisonline which may have caused the data to do this?

Kind Regards,

Roy Langman

GIS Development | ICT services

50 Macquarie Street, Hobart, Tasmania, Australia, 7000 | hobartcity.com.au<http://www.hobartcity.com.au/>

Telephone (03) 6238 2452

0 Kudos
HCCGISICT_
New Contributor III

Hello Paul,

Copied the feature classes, deleted all polygons which were smaller than one square metre, republished then added the new feature services back to my original map.  The problem appears to have resolved.

I now need to deal with and issue surrounding the authors of the data set, I don't have permission to make theapparent changes I need in order to support an AGOL map service.

Did you have any luck with the question relating to why this error has occurred given we had not changed the underlying feature service?

Once again, thank you for the response.

Roy.

0 Kudos
PaulTrevillion
Esri Contributor

Glad to learn that the test brought you closer to a resolution. I'm not aware of any changes to the projection algorithms, but I'll see what I can come up with. I gather that your data is unchanged between the time when everything was working fine and when you started to notice display problems with these two layers.

0 Kudos
HCCGISICT_
New Contributor III

Paul you are correct, the two feature services remained unchanged for a period of months.

0 Kudos
ChrisWhitmore
Esri Regular Contributor

Hi Roy,

Would it be possible for you to send the exported shapefile to me at cwhitmore@esri.com? We'll be able to troubleshoot and reproduce the issue internally (data would be kept confidential).

Thanks,

Chris

0 Kudos
HCCGISICT_
New Contributor III

Hello Chris,

I'll send you both shapefiles.

Regards,

Roy Langman

0 Kudos