我在解決導(dǎo)致性能下降的特定情況時遇到了一些麻煩。我很確定這是可以做到的，但我不知道該怎么做。這是暴露問題的示例模式：type Answer{ answerId: String! text: String! topic: Topic!}type Topic { topicId: String! name: String! level: Int!}extend type Query { answer(answerId: String!): Answer! answers: [Answer!]! }我遵循了文檔，特別是這部分https://gqlgen.com/getting-started/#dont-eagerly-fetch-the-user 從我的架構(gòu)中，它生成以下解析器：func (r *queryResolver) Answer(ctx context.Context, answerId string) (*models.Answer, error) {...#Single Query which retrives single record of Answer from DB.#Fills a model Answer with the Id and the text#Proceeds by calling the Topic resolver...}func (r *queryResolver) Answers(ctx context.Context) ([]*models.Answer, error) {...#Single Query which retrives list of Answers from DB#Fills a list of model Answer with the Id and the text-->#For each element of that list, it calls the Topic resolver...}func (r *answerResolver) Topic(ctx context.Context, obj *models.Answer) (*models.Topic, error) {...#Single Query which retrives single record of Topic from DB#Return a model Topic with id, name and level...}當(dāng)使用參數(shù)answer調(diào)用查詢時，將觸發(fā)解析器，它解析屬性并調(diào)用解析器。解析器按預(yù)期工作，檢索并將其合并到并返回。answerIdanswertextTopicTopicTopicAnswer當(dāng)answers查詢在沒有answerId參數(shù)的情況下被調(diào)用時，answer解析器被觸發(fā)，它answers通過單個查詢檢索列表。然后，對于該列表的每個元素，它調(diào)用Topic解析器。Topic 檢索 a并將其Topic合并到單 例中Answer并返回。結(jié)果在這兩種情況下都可以，但是answers如果我要求很多答案，查詢就會成為性能問題。對于每個答案，Topic解析器都會被觸發(fā)并執(zhí)行查詢以檢索單個記錄。前任。如果我有 2 個答案 --> 1 個查詢[Answer0, Answer1]，然后 1 個查詢 Topic0和 1 個Topic1前任。10 個答案 --> 1 個[Answer0, ..., Answer9]，然后每個 10 個TopicN我想獲得一些topic數(shù)組解析器，例如func (r *answersResolver) Topics(ctx context.Context, obj *[]models.Answer) (*[]models.Topic, error) {...#Single Query which retrives list of Topics from DB#Return a list of model Topic with id, name and level...}我希望返回數(shù)組的每個元素都與數(shù)組的相應(yīng)元素合并Answers。有可能嗎？我在哪里可以找到這種方法的例子？謝謝
1 回答
慕蓋茨4494581
TA貢獻(xiàn)1850條經(jīng)驗 獲得超11個贊
可以使用數(shù)據(jù)加載器（文檔）解決問題
我必須為以下數(shù)據(jù)源實現(xiàn)Topics：
package dataloader
import (
"github.com/graph-gophers/dataloader"
)
type ctxKey string
const (
loadersKey = ctxKey("dataloaders")
)
type TopicReader struct {
conn *sql.DB
}
func (t *TopicReader) GetTopics(ctx context.Context, keys dataloader.Keys) []*dataloader.Result {
topicIDs := make([]string, len(keys))
for ix, key := range keys {
topicIDs[ix] = key.String()
}
res := u.db.Exec(
r.Conn,
"SELECT id, name, level
FROM topics
WHERE id IN (?" + strings.Repeat(",?", len(topicIDs-1)) + ")",
topicIDs...,
)
defer res.Close()
output := make([]*dataloader.Result, len(keys))
for index, _ := range keys {
output[index] = &dataloader.Result{Data: res[index], Error: nil}
}
return output
}
type Loaders struct {
TopicLoader *dataloader.Loader
}
func NewLoaders(conn *sql.DB) *Loaders {
topicReader := &TopicReader{conn: conn}
loaders := &Loaders{
TopicLoader: dataloader.NewBatchedLoader(t.GetTopics),
}
return loaders
}
func Middleware(loaders *Loaders, next http.Handler) http.Handler {
return http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
nextCtx := context.WithValue(r.Context(), loadersKey, loaders)
r = r.WithContext(nextCtx)
next.ServeHTTP(w, r)
})
}
func For(ctx context.Context) *Loaders {
return ctx.Value(loadersKey).(*Loaders)
}
func GetTopic(ctx context.Context, topicID string) (*model.Topic, error) {
loaders := For(ctx)
thunk := loaders.TopicLoader.Load(ctx, dataloader.StringKey(topicID))
result, err := thunk()
if err != nil {
return nil, err
}
return result.(*model.Topic), nil
}
- 1 回答
- 0 關(guān)注
- 137 瀏覽
添加回答
舉報
0/150
提交
取消