1

I'm am using a struct Contact that has a Load() method. The idea is that Load() populates Contact's fields with data. The data is unmarshalled from json returned by a server. Once loaded, I want to be able to access the data on Contact.

I know that the json is unmarshalling correctly because I can print the data to the console within Load(), immediately after it has been unmarshalled. However, this data is not present on Contact once it has been loaded. So when I create a new Contact and call its Load() method, its data fields are still empty when I go to access them.

This is a simplified version of my code (minus imports for brevity):

package main

func main() {
    client := &http.Client{}

    mp := myPackage.NewContact("1234", client)
    if err := mp.Load(); err != nil {
        log.Println(err)
    }

    // prints empty string
    log.Println(mp.EmailOptIn)

}

// myPackage

package myPackage

type Contact struct {
    ID          string
    client      *http.Client
    EmailOptIn  string      `json:"marketing_optin"`
}

func NewContact(ID string, client *http.Client) *Contact {
    return &Contact{
        ID:     ID,
        client: client,
    }
}

func (contact Contact) Load() error {
    req, err := http.NewRequest("GET", "https://anAPI.com/"+contact.ID, nil)
    if err != nil {
        log.Println(err)
    }

    resp, err := contact.client.Do(req)
    if err != nil {
        log.Println(err)
    }
    defer resp.Body.Close()

    if resp.StatusCode == 200 {
        body, _ := ioutil.ReadAll(resp.Body)

        if err = json.Unmarshal(body, &contact); err != nil {
            log.Println("Can't unmarshall: " + err.Error())
        }

        // prints "Opted_in"
        log.Println(contact.EmailOptIn)

        return nil
    }

    return errors.New("oh dear")
}
DavidR
  • 359
  • 1
  • 4
  • 15

1 Answers1

7

Contact.Load() modifies the struct, so in order to retain the changes, it must have a pointer receiver:

func (contact *Contact) Load() error {
    // ...
}

All parameters (including the receiver) are passed by value which means a copy is made and the functions / methods can only operate on the copy. If the receiver is not a pointer then all changes the method makes are on the copy which is discarded when the method returns.

If the receiver is a pointer, it will also be copied, but you don't want to modify the pointer but the pointed value, which will be the same.

See related / possible duplicates:

My object is not updated even if I use the pointer to a type to update it

Remove an element of a slice in a struct

daplho
  • 1,113
  • 1
  • 11
  • 25
icza
  • 389,944
  • 63
  • 907
  • 827
  • Awesome. I couldn't figure out before why people sometimes used pointers in receivers and other times not. – DavidR Feb 05 '19 at 10:40