Salesforce is a very flexible platform in that it provides the Metadata API for users to create, read, update and delete their entire Salesforce environment from objects to page layouts and more. This makes it very easy to programmatically setup and teardown the Salesforce environment.

Retrieving Object Metadata

One common use case for the Metadata API is retrieving information about an object (fields, permissions, etc.). First, load the {salesforcer}, {dplyr}, and {purrr} packages and login, if needed.

library(dplyr, warn.conflicts = FALSE)
library(purrr)
library(salesforcer)
sf_auth()

You can use the sf_read_metadata() function to return a list of objects and their metadata. In the example below we retrieve the metadata for the Account and Contact objects. Note that the metadata_type argument is “CustomObject”. Standard Objects are an implementation of CustomObjects, so they are returned using that metadata type.

read_obj_result <- sf_read_metadata(metadata_type = "CustomObject",
                                    object_names = c("Account", "Contact"))

read_obj_result[[1]][c("fullName", "label", "sharingModel", "enableHistory")]
#> $fullName
#> [1] "Account"
#> 
#> $label
#> [1] "Account"
#> 
#> $sharingModel
#> [1] "ReadWrite"
#> 
#> $enableHistory
#> [1] "false"

first_two_fields_idx <- head(which(names(read_obj_result[[1]]) == "fields"), 2)

# show the first two returned fields of the Account object
read_obj_result[[1]][first_two_fields_idx]
#> $fields
#> $fields$fullName
#> [1] "AccountNumber"
#> 
#> $fields$trackFeedHistory
#> [1] "false"
#> 
#> 
#> $fields
#> $fields$fullName
#> [1] "AccountSource"
#> 
#> $fields$trackFeedHistory
#> [1] "false"
#> 
#> $fields$type
#> [1] "Picklist"

The data is returned as a list because object definitions are highly nested representations. You may notice that we are missing some really specific details, such as, the picklist values of a field with type “Picklist”. You can get that information using
sf_describe_object_fields(). Here is an example using sf_describe_object_fields() where we get a tbl_df with one row for each field on the Account object:

acct_fields <- sf_describe_object_fields("Account")
acct_fields %>% select(name, label, length, soapType, type)
#> # A tibble: 68 x 5
#>   name           label            length soapType    type     
#>   <chr>          <chr>            <chr>  <chr>       <chr>    
#> 1 Id             Account ID       18     tns:ID      id       
#> 2 IsDeleted      Deleted          0      xsd:boolean boolean  
#> 3 MasterRecordId Master Record ID 18     tns:ID      reference
#> 4 Name           Account Name     255    xsd:string  string   
#> 5 Type           Account Type     40     xsd:string  picklist 
#> # … with 63 more rows

# show the picklist selection options for the Account Type field
acct_fields %>% 
  filter(label == "Account Type") %>% 
  .$picklistValues
#> [[1]]
#> # A tibble: 7 x 4
#>   active defaultValue label                      value                     
#>   <chr>  <chr>        <chr>                      <chr>                     
#> 1 true   false        Prospect                   Prospect                  
#> 2 true   false        Customer - Direct          Customer - Direct         
#> 3 true   false        Customer - Channel         Customer - Channel        
#> 4 true   false        Channel Partner / Reseller Channel Partner / Reseller
#> 5 true   false        Installation Partner       Installation Partner      
#> # … with 2 more rows

If you prefer to be more precise about collecting and formatting the field data you can work directly with the nested lists that the APIs return. In this example we look at the picklist values of fields on the Account object.

describe_obj_result <- sf_describe_objects(object_names=c('Account', 'Contact'))

# confirm that the Account object is queryable
describe_obj_result[[1]][c('label', 'queryable')]
#> $label
#> [1] "Account"
#> 
#> $queryable
#> [1] "true"

# now show the different picklist values for the Account Type field
all_fields <- describe_obj_result[[1]][names(describe_obj_result[[1]]) == "fields"]

acct_type_field_idx <- which(sapply(all_fields, 
                                    FUN=function(x){x$label}) == "Account Type")
acct_type_field <- all_fields[[acct_type_field_idx]]
acct_type_field[which(names(acct_type_field) == "picklistValues")] %>% 
  map_df(as_tibble)
#> # A tibble: 7 x 4
#>   active defaultValue label                      value                     
#>   <chr>  <chr>        <chr>                      <chr>                     
#> 1 true   false        Prospect                   Prospect                  
#> 2 true   false        Customer - Direct          Customer - Direct         
#> 3 true   false        Customer - Channel         Customer - Channel        
#> 4 true   false        Channel Partner / Reseller Channel Partner / Reseller
#> 5 true   false        Installation Partner       Installation Partner      
#> # … with 2 more rows

It is recommended that you try out the various metadata functions sf_read_metadata(), sf_list_metadata(), sf_describe_metadata(), and sf_describe_objects() in order to see which information best suits your use case.

Modifying Metadata

Where the Metadata API really shines is when it comes to CRUD operations on metadata. In this example we will create an object, add fields to it, then delete that object.

# create an object
base_obj_name <- "Custom_Account1"
custom_object <- list()
custom_object$fullName <- paste0(base_obj_name, "__c")
custom_object$label <- paste0(gsub("_", " ", base_obj_name))
custom_object$pluralLabel <- paste0(base_obj_name, "s")
custom_object$nameField <- list(displayFormat = 'AN-{0000}', 
                                label = paste0(base_obj_name, ' Number'), 
                                type = 'AutoNumber')
custom_object$deploymentStatus <- 'Deployed'
custom_object$sharingModel <- 'ReadWrite'
custom_object$enableActivities <- 'true'
custom_object$description <- paste0(base_obj_name, " created by the Metadata API")
custom_object_result <- sf_create_metadata(metadata_type = 'CustomObject',
                                           metadata = custom_object)

# add fields to the object
custom_fields <- tibble(fullName=c(paste0(base_obj_name, '__c.CustomField3__c'), 
                                   paste0(base_obj_name, '__c.CustomField4__c')), 
                        label=c('Test Field3', 'Test Field4'), 
                        length=c(100, 100), 
                        type=c('Text', 'Text'))
create_fields_result <- sf_create_metadata(metadata_type = 'CustomField', 
                                           metadata = custom_fields)

# delete the object
delete_obj_result <- sf_delete_metadata(metadata_type = 'CustomObject', 
                                        object_names = c('Custom_Account1__c'))

Note that newly created custom fields are not editable by default, meaning that you will not be able to insert records into them until updating the field level security of your user profile. Run the following code to determine the user profiles in your org and updating the field permissions on an object that you may have created with the example code above.

# get list of user proviles in order to get the "fullName" 
# parameter correct in the next call
my_queries <- list(list(type = 'Profile'))
profiles_list <- sf_list_metadata(queries = my_queries)

admin_editable <- list(fullName = 'Admin', 
                       fieldPermissions=list(field=paste0(custom_object$fullName, 
                                                          '.CustomField3__c'),
                                             editable='true'), 
                       fieldPermissions=list(field=paste0(custom_object$fullName, 
                                                          '.CustomField4__c'),
                                             editable='true'))

# update the field level security to "editable" for your fields
prof_update <- sf_update_metadata(metadata_type = 'Profile', 
                                  metadata = admin_editable)

# now try inserting values into that custom object's fields
my_new_data = tibble(CustomField3__c = "Hello World", CustomField4__c = "Hello World")
added_record <- sf_create(my_new_data, object_name = custom_object$fullName)