Swift's Encodable
/Decodable
protocols, released with Swift 4, make JSON (de)serialization quite pleasant. However, I have not yet found a way to have fine-grained control over which properties should be encoded and which should get decoded.
I have noticed that excluding the property from the accompanying CodingKeys
enum excludes the property from the process altogether, but is there a way to have more fine-grained control?
The list of keys to encode/decode is controlled by a type called CodingKeys
(note the s
at the end). The compiler can synthesize this for you but can always override that.
Let's say you want to exclude the property nickname
from both encoding and decoding:
struct Person: Codable {
var firstName: String
var lastName: String
var nickname: String?
private enum CodingKeys: String, CodingKey {
case firstName, lastName
}
}
If you want it to be asymmetric (i.e. encode but not decode or vice versa), you have to provide your own implementations of encode(with encoder: )
and init(from decoder: )
:
struct Person: Codable {
var firstName: String
var lastName: String
// Since fullName is a computed property, it's excluded by default
var fullName: String {
return firstName + " " + lastName
}
private enum CodingKeys: String, CodingKey {
case firstName, lastName, fullName
}
// We don't want to decode `fullName` from the JSON
init(from decoder: Decoder) throws {
let container = try decoder.container(keyedBy: CodingKeys.self)
firstName = try container.decode(String.self, forKey: .firstName)
lastName = try container.decode(String.self, forKey: .lastName)
}
// But we want to store `fullName` in the JSON anyhow
func encode(to encoder: Encoder) throws {
var container = encoder.container(keyedBy: CodingKeys.self)
try container.encode(firstName, forKey: .firstName)
try container.encode(lastName, forKey: .lastName)
try container.encode(fullName, forKey: .fullName)
}
}