如何在Go中解析JSON？

Question

JSON

{
  "id" : "12387",
  "inv" :[
    {
      "qty" : 5,

       "seq" : 2,
       "invIs" : "1HG9876",
       "addCharges" :[
         {
          "amnt" : 24,
          "char" : "REI",
          "type" : "MT"
          },
          {
          "amnt" : 24,
          "char" : "REI",
          "type" : "MT"
          }
        ],

      "seq" : 3

    },
    {
      "qty" : 5,

       "seq" : 2,
       "invIs" : "1HG9876",
       "addCharges" :[
         {
          "amnt" : 24,
          "char" : "REI",
          "type" : "MT"
          },
          {
          "amnt" : 24,
          "char" : "REI",
          "type" : "MT"
          }
        ],

      "seq" : 3

    }
  ],
    "charges" : {
      "fee" : 24 ,
      "bkg" : 7676
    }

}

My JSON structure is like that shown above. I need to take amnt in inv-addCharges in an array. If it has ten elements in the array, I need to get that in an array containing the particular amnt alone in such a way like

[{"amnt" : 34 } ,{"amnt" : 34} .... so on ]

Things I tried:

var j map[string]interface{}
    err := json.Unmarshal([]byte(ticket), &j)
    if err != nil {
        panic(err)
    }

    // Pull out the parents object
    bytInv := j["inv"].([]interface{})    

    // // Print out mother and father
    fmt.Println(bytInv)

Output

[map[qty:5 seq:3 invIs:1HG9876 addCharges:[map[amnt:24 char:REI type:MT] map[amnt:24 char:REI type:MT]]] map[qty:5 seq:3 invIs:1HG9876 addCharges:[map[amnt:24 char:REI type:MT] map[amnt:24 char:REI type:MT]]]]

After this, I was not able to proceed any further.

Note: I don't want to use structs because I have many JSON structures for this. And that's my requirement given.

Answer 1

Requirement or not, I really don't see why you insist on not using structs for this. Can you elaborate on exactly why this is the case?

As evidenced by the example program below, the solution is so much easier to understand and reason about if you just represent the data as properly typed structs. You can run it on the Go playground to see it in action.

There is one part of your JSON data which confuses me. Each Item node contains the seq field twice. Both with a different value. Which ever way you use to interpret the data, one of the fields will be lost. Depending on the implementation of the unmarshaller, it will either ignore the second field, or overwrite the first one with the second value. This is a bug in the way your data is generated.

package main

import (
    "encoding/json"
    "fmt"
    "os"
)

func main() {
    charges, err := findCharges()
    if err != nil {
        fmt.Fprintf(os.Stderr, "%v
", err)
        os.Exit(1)
    }

    for _, c := range charges {
        fmt.Printf("%+v
", c)
    }
}

// findCharges locates all 'AddCharge` instances and returns them in a slice.
func findCharges() ([]AddCharge, error) {
    var prod Product

    data := []byte(jsonString)
    err := json.Unmarshal(data, &prod)
    if err != nil {
        return nil, err
    }

    var charges []AddCharge

    for _, item := range prod.Items {
        charges = append(charges, item.AddCharges...)
    }

    return charges, nil
}

type Product struct {
    Id    string `json:"id"`
    Items []Item `json:"inv"`
}

type Item struct {
    Quantity   int         `json:"qty"`
    Sequence   int         `json:"seq"`
    Inventory  string      `json:"invIs"`
    AddCharges []AddCharge `json:"addCharges"`
    Charges    []Charge    `json:"charges"`
}

type Charge struct {
    Fee int `json:"fee"`
    Bkg int `json:"bkg"`
}

type AddCharge struct {
    Amount int    `json:"amnt"`
    Char   string `json:"char"`
    Type   string `json:"type"`
}

const jsonString = `{
  "id" : "12387",
  "inv" :[
    {
      "qty" : 5,
       "seq" : 2,
       "invIs" : "1HG9876",
       "addCharges" :[
         {
          "amnt" : 24,
          "char" : "REI",
          "type" : "MT"
          },
          {
          "amnt" : 12,
          "char" : "REI",
          "type" : "MT"
          }
        ],

      "seq" : 3

    },
    {
      "qty" : 5,

       "seq" : 2,
       "invIs" : "1HG9876",
       "addCharges" :[
         {
          "amnt" : 64,
          "char" : "REI",
          "type" : "MT"
          },
          {
          "amnt" : 36,
          "char" : "REI",
          "type" : "MT"
          }
        ],

      "seq" : 3

    }
  ],
    "charges" : {
      "fee" : 24 ,
      "bkg" : 7676
    }

}`

edit: If the addCharges data is really all you are interested in, you can omit all the fields you do not need from the structs. The unmarshaller will simply ignore the data for which the struct has no field defined.