Pagination access #64

Activity

1. aren55555 commented on Jan 25, 2017

   

   aren55555

   on Jan 25, 2017

   Contributor

   More actions

   @gopherfortress we recently added features to easily add "links" when serializing/marshaling. Unfortunately the same cannot be said for deserializing/unmarshaling.

   However there is still a way you can parse them out:

   func TestManyPayload_withLinks(t *testing.T) {
   	firstPageURL := "http://somesite.com/movies?page[limit]=50&page[offset]=50"
   	prevPageURL := "http://somesite.com/movies?page[limit]=50&page[offset]=0"
   	nextPageURL := "http://somesite.com/movies?page[limit]=50&page[offset]=100"
   	lastPageURL := "http://somesite.com/movies?page[limit]=50&page[offset]=500"
   
   	sample := map[string]interface{}{
   		"data": []interface{}{
   			map[string]interface{}{
   				"type": "posts",
   				"id":   "1",
   				"attributes": map[string]interface{}{
   					"body":  "First",
   					"title": "Post",
   				},
   			},
   			map[string]interface{}{
   				"type": "posts",
   				"id":   "2",
   				"attributes": map[string]interface{}{
   					"body":  "Second",
   					"title": "Post",
   				},
   			},
   		},
   		"links": map[string]interface{}{
   			KeyFirstPage:    firstPageURL,
   			KeyPreviousPage: prevPageURL,
   			KeyNextPage:     nextPageURL,
   			KeyLastPage:     lastPageURL,
   		},
   	}
   
   	data, err := json.Marshal(sample)
   	if err != nil {
   		t.Fatal(err)
   	}
   	in := bytes.NewReader(data)
   
   	payload := new(ManyPayload)
   	if err = json.NewDecoder(in).Decode(payload); err != nil {
   		t.Fatal(err)
   	}
   
   	if payload.Links == nil {
   		t.Fatal("Was expecting a non nil ptr Link field")
   	}
   
   	links := *payload.Links
   
   	first, ok := links[KeyFirstPage]
   	if !ok {
   		t.Fatal("Was expecting a non nil ptr Link field")
   	}
   	if e, a := firstPageURL, first; e != a {
   		t.Fatalf("Was expecting links.%s to have a value of %s, got %s", KeyFirstPage, e, a)
   	}
   }
2. aren55555

   closed this as completedon Jan 25, 2017
3. gopherfortress commented on Jan 25, 2017

   

   gopherfortress

   on Jan 25, 2017

   Author

   More actions

   Thanks for the answer!

   Any idea of the amount of work that is being done when doing the

   payload := new(ManyPayload)
   if err = json.NewDecoder(in).Decode(payload); err != nil {
   	t.Fatal(err)
   }
   

   My concern is that since links are at the very bottom of the document then it is possible that, assuming a big JSON document, twice the work will be done being done, once for decoding the pagination links and once for decoding the rest.

   Unfortunately the same cannot be said for deserializing/unmarshaling.

   Assuming there is indeed twice the work being done, any plans to support unmarshaling of pagination links?

   Thank you!
4. aren55555 commented on Jan 25, 2017

   

   aren55555

   on Jan 25, 2017

   Contributor

   More actions

   @gopherfortress if all you are after is the "links" member there is definitely a more performant solution. The code sample I posted will traverse all of the "data" member which could be fairly lengthy. Take a look at our source code; I think you want to start by altering the OnePayload or ManyPayload to fit your use case of only parsing "links".

   As for fully supporting unmarshalling "links" that is a goal - however initial investigation leads me to believe adding such a feature (in a simple way) would mean breaking the existing unmarshal method signatures. We still need to put more thought into this.