This tutorial demonstrates how to add new fields to an existing feature service using R. Adding fields allows you to expand your published data without creating an entirely new service.
Prerequisites and Permissions
To add fields to a feature service, you must be either the service owner or have administrative privileges. This is a potentially destructive operation, so proceed with caution and consider testing on non-production data first.
Common Use Cases
Adding fields to existing feature services is useful when you need to:
- Extend a published feature service with additional data columns
- Add model predictions or calculated scores to existing spatial data
- Incorporate new attributes discovered after initial publication
- Append analytical results without republishing the entire dataset
Authentication Setup
Before beginning, you’ll need to authenticate to ArcGIS Online or ArcGIS Enterprise. For detailed authentication approaches, see the authentication guide.
This tutorial uses the Palmer Penguins dataset as an example. We’ll initially publish the data without the body_mass field, then demonstrate how to add it later. For more comprehensive publishing guidance, refer to the publishing tutorial.
library(arcgis) set_arc_token(auth_user())
Publishing the Initial Dataset
We’ll start by publishing the penguins dataset with all columns except body_mass, which we’ll add later to demonstrate the field addition process:
# create a random unique name for publishing layer_name <- paste0("penguins-", ulid::ulid()) # subset to all but body mass to_publish <- subset(penguins, select = -body_mass) # publish a layer published <- publish_layer(to_publish, layer_name)
new_service_url <- published$services$encodedServiceURL penguins_fl <- arc_open(new_service_url) |> get_layer(0) penguins_fl
#> <Table> #> Name: penguins-01K6KAXQCKH60AT3EBP6768YT5 #> Capabilities: Create,Delete,Query,Update,Editing
Adding New Fields to the Service Definition
Now we’ll add the missing body_mass field to our published feature service. The service definition contains metadata that describes the service structure, including fields, symbology, and indexes.
Defining Field Types
First, we need to define the field structure for the new column. The arcgisutils::as_fields() function automatically infers appropriate field types from your R data:
field_types <- as_fields(penguins[,"body_mass", drop = FALSE]) field_types
#> name type #> integer body_mass esriFieldTypeInteger #> alias length nullable #> integer body_mass NA TRUE #> editable #> integer TRUE
Important: Only include the new fields you want to add. Including existing field names will cause an error since fields cannot be duplicated in the service definition.
Updating the Service Definition
Use arcgislayers::add_layer_definition() to add the new field to the feature service. This function updates the service schema and returns the modified feature layer:
# update the feature service definition penguins_fl <- add_layer_definition( penguins_fl, fields = field_types )
Verifying the New Field
Let’s confirm that our new field has been successfully added to the feature service:
list_fields(penguins_fl)
#> # A data frame: 9 × 10 #> name type actualType alias sqlType #> * <chr> <chr> <chr> <chr> <chr> #> 1 objec… esri… int obje… sqlTyp… #> 2 speci… esri… nvarchar spec… sqlTyp… #> 3 island esri… nvarchar isla… sqlTyp… #> 4 bill_… esri… float bill… sqlTyp… #> 5 bill_… esri… float bill… sqlTyp… #> 6 flipp… esri… int flip… sqlTyp… #> 7 sex esri… nvarchar sex sqlTyp… #> 8 year esri… int year sqlTyp… #> 9 body_… esri… <NA> body… sqlTyp… #> # ℹ 5 more variables: nullable <lgl>, #> # editable <lgl>, domain <lgl>, #> # defaultValue <lgl>, length <int>
You should now see the body_mass field listed among the service fields.
Next Steps
With the new field added to the service definition, you can now populate it with data using arcgislayers::update_features(). The field structure is in place and ready to receive values that match the defined field type.
This approach allows you to iteratively expand your feature services as your data requirements evolve, without the need to republish entirely new services.