我正在为我用 go 编写的 be 的 graphql 查询开发一个解析器函数。在解析器中,我有想要更新的用户数据,使用包含多个可能的更新属性的输入值。
在 javascript 中,这可以通过解构(伪)快速完成:
const mergedobj = {...oldprops, ...newprops}
目前,我的解析器函数如下所示(使用 gqlgen 作为 graphql go 解析器):
func (r *mutationResolver) ModifyUser(ctx context.Context, input *model.ModifyUserInput) (*model.User, error) {
id := input.ID
us, ok := r.Resolver.UserStore[id]
if !ok {
return nil, fmt.Errorf("not found")
}
if input.FirstName != nil {
us.FirstName = *input.FirstName
}
if input.LastName != nil {
us.LastName = *input.LastName
}
if input.ProfileImage != nil {
us.ProfileImage = input.ProfileImage
}
if input.Password != nil {
us.Password = *input.Password
}
if input.Email != nil {
us.Email = *input.Email
}
if input.InTomorrow != nil {
us.InTomorrow = input.InTomorrow
}
if input.DefaultDaysIn != nil {
us.DefaultDaysIn = input.DefaultDaysIn
}
r.Resolver.UserStore[id] = us
return &us, nil
}
这感觉很简单。在这种情况下迭代结构键有意义吗?或者我还缺少另一种模式吗?
正确答案
使用函数来减少样板:
func mergef[t any](a, b *t) {
if b != nil {
*a = *b
}
}
...
mergef(&us.firstname, input.firstname)
mergef(&us.lastname, input.lastname)
...
使用反射包来减少更多样板:
// merge sets fields in struct pointed to by d to
// dereferenced fields in struct pointed to by s.
//
// argument s must point to a struct with pointer type
// fields.
// argument d must point to a struct with fields that
// correspond to the fields in s: there must be a field
// in d with the same name as a field in s; the type of
// the field in s must be a pointer to the type of the field
// in d.
func merge(d, s any) {
sv := reflect.valueof(s).elem()
dv := reflect.valueof(d).elem()
for i := 0; i < sv.numfield(); i++ {
sf := sv.field(i)
if sf.isnil() {
continue
}
df := dv.fieldbyname(sv.type().field(i).name)
df.set(sf.elem())
}
}
使用这样的函数:
merge(us, input)