I have the following type in Golang:
type Base64Data []byte
In order to support unmarshalling a base64 encoded string to this type, I did the following:
func (b *Base64Data) UnmarshalJSON(data []byte) error {
if len(data) == 0 {
return nil
}
content, err := base64.StdEncoding.DecodeString(string(data[1 : len(data)-1]))
if err != nil {
return err
}
*b = []byte(xml)
return nil
}
Now I also want to be able to marshal and unmarshal it to mongo database, using mgo Golang library.
The problem is that I already have documents there stored as base64 encoded string, so I have to maintain that.
I tried to do the following:
func (b Base64Data) GetBSON() (interface{}, error) {
return base64.StdEncoding.EncodeToString([]byte(b)), nil
}
func (b *Base64DecodedXml) SetBSON(raw bson.Raw) error {
var s string
var err error
if err = raw.Unmarshal(&s); err != nil {
return err
}
*b, err = base64.StdEncoding.DecodeString(s)
return err
}
So that after unmarshaling, the data is already decoded, so I need to encode it back, and return it as a string so it will be written to db as a string (and vice versa)
For that I implemented bson getter and setter, but it seems only the getter is working properly
JSON unmarshaling from base64 encoded string works, as well marshaling it to database. but unmarshling setter seems to not be called at all.
Can anyone suggest what I'm missing, so that I'll be able to properly hold the data decoded in memory, but encoded string type?
This is a test I tried to run:
b := struct {
Value shared.Base64Data `json:"value" bson:"value"`
}{}
s := `{"value": "PHJvb3Q+aGVsbG88L3Jvb3Q+"}`
require.NoError(t, json.Unmarshal([]byte(s), &b))
t.Logf("%v", string(b.Value))
b4, err := bson.Marshal(b)
require.NoError(t, err)
t.Logf("%v", string(b4))
require.NoError(t, bson.Unmarshal(b4, &b))
t.Logf("%v", string(b.Value))
You can't marshal any value with bson.Marshal(), only maps and struct values.
If you want to test it, pass a map, e.g. bson.M to bson.Marshal():
var x = Base64Data{0x01, 0x02, 0x03}
dd, err := bson.Marshal(bson.M{"data": x})
fmt.Println(string(dd), err)
Your code works as-is, and as you intend it to. Try to insert a wrapper value to verify it:
c := sess.DB("testdb").C("testcoll")
var x = Base64Data{0x01, 0x02, 0x03}
if err := c.Insert(bson.M{
"data": x,
}); err != nil {
panic(err)
}
This will save the data as a string, being the Base64 encoded form.
Of course if you want to load it back into a value of type Base64Data, you will also need to define the SetBSON(raw Raw) error method too (bson.Setter interface).
Related
I thought I understood unmarshalling by now, but I guess not. I'm having a little bit of trouble unmarshalling a map in go. Here is the code that I have so far
type OHLC_RESS struct {
Pair map[string][]Candles
Last int64 `json:"last"`
}
type Candles struct {
Time uint64
Open string
High string
Low string
Close string
VWAP string
Volume string
Count int
}
func (c *Candles) UnmarshalJSON(d []byte) error {
tmp := []interface{}{&c.Time, &c.Open, &c.High, &c.Low, &c.Close, &c.VWAP, &c.Volume, &c.Count}
length := len(tmp)
err := json.Unmarshal(d, &tmp)
if err != nil {
return err
}
g := len(tmp)
if g != length {
return fmt.Errorf("Lengths don't match: %d != %d", g, length)
}
return nil
}
func main() {
response := []byte(`{"XXBTZUSD":[[1616662740,"52591.9","52599.9","52591.8","52599.9","52599.1","0.11091626",5],[1616662740,"52591.9","52599.9","52591.8","52599.9","52599.1","0.11091626",5]],"last":15}`)
var resp OHLC_RESS
err := json.Unmarshal(response, &resp)
fmt.Println("resp: ", resp)
}
after running the code, the last field will unmarshal fine, but for whatever reason, the map is left without any value. Any help?
The expedient solution, for the specific example JSON, would be to NOT use a map at all but instead change the structure of OHLC_RESS so that it matches the structure of the JSON, i.e.
type OHLC_RESS struct {
Pair []Candles `json:"XXBTZUSD"`
Last int64 `json:"last"`
}
https://go.dev/play/p/Z9PhJt3wX33
However it's safe to assume, I think, that the reason you've opted to use a map is because the JSON object's key(s) that hold the "pairs" can vary and so hardcoding them into the field's tag is out of the question.
To understand why your code doesn't produce the desired result, you have to realize two things. First, the order of a struct's fields has no bearing on how the keys of a JSON object will be decoded. Second, the name Pair holds no special meaning for the unmarshaler. Therefore, by default, the unmarshaler has no way of knowing that your wish is to decode the "XXBTZUSD": [ ... ] element into the Pair map.
So, to get your desired result, you can have the OHLC_RESS implement the json.Unmarshaler interface and do the following:
func (r *OHLC_RESS) UnmarshalJSON(d []byte) error {
// first, decode just the object's keys and leave
// the values as raw, non-decoded JSON
var obj map[string]json.RawMessage
if err := json.Unmarshal(d, &obj); err != nil {
return err
}
// next, look up the "last" element's raw, non-decoded value
// and, if it is present, then decode it into the Last field
if last, ok := obj["last"]; ok {
if err := json.Unmarshal(last, &r.Last); err != nil {
return err
}
// remove the element so it's not in
// the way when decoding the rest below
delete(obj, "last")
}
// finally, decode the rest of the element values
// in the object and store them in the Pair field
r.Pair = make(map[string][]Candles, len(obj))
for key, val := range obj {
cc := []Candles{}
if err := json.Unmarshal(val, &cc); err != nil {
return err
}
r.Pair[key] = cc
}
return nil
}
https://go.dev/play/p/Lj8a8Gx9fWH
I need to parse this json
{
"version": "1.1.29-snapshot",
"linux-amd64": {
"url": "https://origin/path",
"size": 7794688,
"sha256": "14b3c3ad05e3a98d30ee7e774646aec7ffa8825a1f6f4d9c01e08bf2d8a08646"
},
"windows-amd64": {
"url": "https://origin/path",
"size": 8102400,
"sha256": "01b8b927388f774bdda4b5394e381beb592d8ef0ceed69324d1d42f6605ab56d"
}
}
Keys like linux-amd64 are dynamic and theirs amount is arbitrary. I tried something like that to describe it and unmarshal. Obviously it doesn't work. Items is always empty.
type FileInfo struct {
Url string `json:"url"`
Size int64 `json:"size"`
Sha256 string `json:"sha256"`
}
type UpdateInfo struct {
Version string `json:"version"`
Items map[string]FileInfo
}
It's similar to this use case, but has no parent key items. I suppose I can use 3rd party library or map[string]interface{} approach, but I'm interested in knowing how to achieve this with explicitly declared types.
The rest of the parsing code is:
func parseUpdateJson(jsonStr []byte) (UpdateInfo, error) {
var allInfo = UpdateInfo{Items: make(map[string]FileInfo)}
var err = json.Unmarshal(jsonStr, &allInfo)
return allInfo, err
}
Look at the link I attached and you will realize that is not that simple as you think. Also I pointed that I interested in typed approach. Ok, how to declare this map[string]FileInfo to get parsed?
You can create a json.Unmarshaller to decode the json into a map, then apply those values to your struct: https://play.golang.org/p/j1JXMpc4Q9u
type FileInfo struct {
Url string `json:"url"`
Size int64 `json:"size"`
Sha256 string `json:"sha256"`
}
type UpdateInfo struct {
Version string `json:"version"`
Items map[string]FileInfo
}
func (i *UpdateInfo) UnmarshalJSON(d []byte) error {
tmp := map[string]json.RawMessage{}
err := json.Unmarshal(d, &tmp)
if err != nil {
return err
}
err = json.Unmarshal(tmp["version"], &i.Version)
if err != nil {
return err
}
delete(tmp, "version")
i.Items = map[string]FileInfo{}
for k, v := range tmp {
var item FileInfo
err := json.Unmarshal(v, &item)
if err != nil {
return err
}
i.Items[k] = item
}
return nil
}
This answer is adapted from this recipe in my YouTube video on advanced JSON handling in Go.
func (u *UpdateInfo) UnmarshalJSON(d []byte) error {
var x struct {
UpdateInfo
UnmarshalJSON struct{}
}
if err := json.Unmarshal(d, &x); err != nil {
return err
}
var y map[string]json.RawMessage{}
if err := json.Unsmarshal(d, &y); err != nil {
return err
}
delete(y, "version"_ // We don't need this in the map
*u = x.UpdateInfo
u.Items = make(map[string]FileInfo, len(y))
for k, v := range y {
var info FileInfo
if err := json.Unmarshal(v, &info); err != nil {
return err
}
u.Items[k] = info
}
return nil
}
It:
Unmarshals the JSON into the struct directly, to get the struct fields.
It re-unmarshals into a map of map[string]json.RawMessage to get the arbitrary keys. This is necessary since the value of version is not of type FileInfo, and trying to unmarshal directly into map[string]FileInfo will thus error.
It deletes the keys we know we already got in the struct fields.
It then iterates through the map of string to json.RawMessage, and finally unmarshals each value into the FileInfo type, and stores it in the final object.
If you really don't want to unmarshal multiple times, your next best option is to iterate over the JSON tokens in your input by using the json.Decoder type. I've done this in a couple of performance-sensitive bits of code, but it makes your code INCREDIBLY hard to read, and in almost all cases is not worth the effort.
I need to parse really long json file (more than million items). I don't want to load it to the memory and read it chunk by chunk. There's a good example with the array of items here. The problem is that I deal with the map. And when I call Decode I get not at beginning of value.
I can't get what should be changed.
const data = `{
"object1": {"name": "cattle","location": "kitchen"},
"object2": {"name": "table","location": "office"}
}`
type ReadObject struct {
Name string `json:"name"`
Location string `json:"location"`
}
func ParseJSON() {
dec := json.NewDecoder(strings.NewReader(data))
tkn, err := dec.Token()
if err != nil {
log.Fatalf("failed to read opening token: %v", err)
}
fmt.Printf("opening token: %v\n", tkn)
objects := make(map[string]*ReadObject)
for dec.More() {
var nextSymbol string
if err := dec.Decode(&nextSymbol); err != nil {
log.Fatalf("failed to parse next symbol: %v", err)
}
nextObject := &ReadObject{}
if err := dec.Decode(&nextObject); err != nil {
log.Fatalf("failed to parse next object")
}
objects[nextSymbol] = nextObject
}
tkn, err = dec.Token()
if err != nil {
log.Fatalf("failed to read closing token: %v", err)
}
fmt.Printf("closing token: %v\n", tkn)
fmt.Printf("OBJECTS: \n%v\n", objects)
}
TL,DR: when you are calling Token() method for a first time, you move offset from the beginning (of a JSON value) and therefore you get the error.
You are working with this struct (link):
type Decoder struct {
// others fields omits for simplicity
tokenState int
}
Pay attention for a tokenState field. This value could be one of (link):
const (
tokenTopValue = iota
tokenArrayStart
tokenArrayValue
tokenArrayComma
tokenObjectStart
tokenObjectKey
tokenObjectColon
tokenObjectValue
tokenObjectComma
)
Let's back to your code. You are calling Token() method. This method obtains first JSON-valid token { and changes tokenState from tokenObjectValue to the tokenObjectStart (link). Now you are "in-an-object" state.
If you try to call Decode() at this point you will get an error (not at beginning of value). This is because allowed states of tokenState for calling Decode() are tokenTopValue, tokenArrayStart, tokenArrayValue, tokenObjectValue, i.e. "full" value, not part of it (link).
To avoid this you can just don't call Token() at all and do something like this:
dec := json.NewDecoder(strings.NewReader(dataMapFromJson))
objects := make(map[string]*ReadObject)
if err := dec.Decode(&objects); err != nil {
log.Fatalf("failed to parse next symbol: %v", err)
}
fmt.Printf("OBJECTS: \n%v\n", objects)
Or, if you want to read chunk-by-chunk, you could keep calling Token() until you reach "full" value. And then call Decode() on this value (I guess this should work).
After consuming the initial { with your first call to dec.Token(), you must :
use dec.Token() to extract the next key
after extracting the key, you can call dec.Decode(&nextObject) to decode an entry
example code :
for dec.More() {
key, err := dec.Token()
if err != nil {
// handle error
}
var val interface{}
err = dec.Decode(&val)
if err != nil {
// handle error
}
fmt.Printf(" %s : %v\n", key, val)
}
https://play.golang.org/p/5r1d8MsNlKb
Golang encoding/json package lets you use ,string struct tag in order to marshal/unmarshal string values (like "309230") into int64 field. Example:
Int64String int64 `json:",string"`
However, this doesn't work for slices, ie. []int64:
Int64Slice []int64 `json:",string"` // Doesn't work.
Is there any way to marshal/unmarshal JSON string arrays into []int64 field?
Quote from https://golang.org/pkg/encoding/json:
The "string" option signals that a field is stored as JSON inside a JSON-encoded string. It applies only to fields of string, floating point, integer, or boolean types. This extra level of encoding is sometimes used when communicating with JavaScript programs:
For anyone interested, I found a solution using a custom type having MarshalJSON() and UnmarshalJSON() methods defined.
type Int64StringSlice []int64
func (slice Int64StringSlice) MarshalJSON() ([]byte, error) {
values := make([]string, len(slice))
for i, value := range []int64(slice) {
values[i] = fmt.Sprintf(`"%v"`, value)
}
return []byte(fmt.Sprintf("[%v]", strings.Join(values, ","))), nil
}
func (slice *Int64StringSlice) UnmarshalJSON(b []byte) error {
// Try array of strings first.
var values []string
err := json.Unmarshal(b, &values)
if err != nil {
// Fall back to array of integers:
var values []int64
if err := json.Unmarshal(b, &values); err != nil {
return err
}
*slice = values
return nil
}
*slice = make([]int64, len(values))
for i, value := range values {
value, err := strconv.ParseInt(value, 10, 64)
if err != nil {
return err
}
(*slice)[i] = value
}
return nil
}
The above solution marshals []int64 into JSON string array. Unmarshaling works from both JSON string and integer arrays, ie.:
{"bars": ["1729382256910270462", "309286902808622", "23"]}
{"bars": [1729382256910270462, 309286902808622, 23]}
See example at https://play.golang.org/p/BOqUBGR3DXm
As you quoted from json.Marshal(), the ,string option only applies to specific types, namely:
The "string" option signals that a field is stored as JSON inside a JSON-encoded string. It applies only to fields of string, floating point, integer, or boolean types.
You want it to work with a slice, but that is not supported by the json package.
If you still want this functionality, you have to write your custom marshaling / unmarshaling logic.
What you presented works, but it is unnecessarily complex. This is because you created your custom logic on slices, but you only want this functionality on individual elements of the slices (arrays). You don't want to change how an array / slice (as a sequence of elements) is rendered or parsed.
So a much simpler solution is to only create a custom "number" type producing this behavior, and elements of slices of this custom type will behave the same.
Our custom number type and the marshaling / unmarshaling logic:
type Int64Str int64
func (i Int64Str) MarshalJSON() ([]byte, error) {
return json.Marshal(strconv.FormatInt(int64(i), 10))
}
func (i *Int64Str) UnmarshalJSON(b []byte) error {
// Try string first
var s string
if err := json.Unmarshal(b, &s); err == nil {
value, err := strconv.ParseInt(s, 10, 64)
if err != nil {
return err
}
*i = Int64Str(value)
return nil
}
// Fallback to number
return json.Unmarshal(b, (*int64)(i))
}
And that's all!
The type using it:
type Foo struct {
Bars []Int64Str `json:"bars"`
}
Testing it the same way as you did yields the same result. Try it on the Go Playground.
In python you can take a json object and grab a specific item from it without declaring a struct, saving to a struct then obtaining the value like in Go. Is there a package or easier way to store a specific value from json in Go?
python
res = res.json()
return res['results'][0]
Go
type Quotes struct {
AskPrice string `json:"ask_price"`
}
quote := new(Quotes)
errJson := json.Unmarshal(content, "e)
if errJson != nil {
return "nil", fmt.Errorf("cannot read json body: %v", errJson)
}
You can decode into a map[string]interface{} and then get the element by key.
func main() {
b := []byte(`{"ask_price":"1.0"}`)
data := make(map[string]interface{})
err := json.Unmarshal(b, &data)
if err != nil {
panic(err)
}
if price, ok := data["ask_price"].(string); ok {
fmt.Println(price)
} else {
panic("wrong type")
}
}
Structs are often preferred as they are more explicit about the type. You only have to declare the fields in the JSON you care about, and you don't need to type assert the values as you would with a map (encoding/json handles that implicitly).
Try either fastjson or jsonparser. jsonparser is optimized for the case when a single JSON field must be selected, while fastjson is optimized for the case when multiple unrelated JSON fields must be selected.
Below is an example code for fastjson:
var p fastjson.Parser
v, err := p.Parse(content)
if err != nil {
log.Fatal(err)
}
// obtain v["ask_price"] as float64
price := v.GetFloat64("ask_price")
// obtain v["results"][0] as generic JSON value
result0 := v.Get("results", "0")