How can I add a shape that has more than 10,000 features to make a map in Online ArcGIS?

Since it is only to show a sample of the work that has been achieved I could cut out information but I don't know how to reduce it to less than 10,000 features.

What I have mapped are census tract areas and food outlets.

I reduced the features by using shapes at the city level instead of the country level but still it is too much information to display.

Any ideas?

  • The limit exists for a purpose: If every user added 10,000+ shapes, it would soon involve petabytes of storage. Looking at your options, it seems an ArcGIS Server feature service is the only way to add that volume of data as features.
    – Vince
    Oct 5, 2015 at 13:11

1 Answer 1


Since this is only for a presentation....

Do some processing in ArcGIS Desktop to solve this problem.
Start by running Simplify Polygon on the dataset (I would suggest BEND_SIMPLIFY since that is better for presentation graphics). The reason for doing this is to improve performance with multipart features or large numbers of features.

Use a tolerance that is appropriate to the scale that you will be showing, e.g. if you are showing 1:100,000 scale and you are okay with a 1/8th inch deviation in your presentation graphic, then use 1041.67 feet for your tolerance. 1/8th" at 1:100,000 -> 0.125"(map) * 100,000 = 12,500"(ground) = 1041.67'.

Next you have some decisions to make. What on the map needs to be clickable?

If you need detailed demographics of every census tract, then you are going to have to do scale dependent aggregation. Your best bet is to create a separate county level layer where you use the Dissolve tool to dissolve tracts at the county level and aggregate the statistics (I bet the census already has this available too). Use this layer at small scales. Set your scale dependent display so that the census tracts themselves only show up at large scales. If you want to do the whole United States, you will want to aggregate again at the state level (though I would suggest just having state outlines and not bothering with clickable statistics).

If you do not need detailed demographics on each census tract and instead just need to display areas (e.g. which tracts have food outlets), Dissolve is your best friend. Ignore boundaries and dissolve together the census tracts with the same characteristics. Allow multipart features. This will create some features with a large number of vertices, but this is why you simplified in the first place. Note that clicking on these features will be meaningless except for the attribute that you dissolve on. (So, if the attribute is just presence or absence of a food outlet, clicking will only show that. If the the attribute is "number of food outlets per tract" then that is all that clicking will show.)

Lastly, for whatever strategy you use you can cut tracts that are not meaningful, or split them out into a new layer. So, if census tracts without food outlets (or maybe with) are not meaningful to your presentation, you can drop them completely from the layer or split them out into a separate layer that will have its own 10,000 count limit.

In short, use multiple layers in meaningful ways. Use scale dependent generalization. Dissolve is your best friend. Make intelligent decisions about what information you can eliminate or split out.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.