dgraph 创建节点和边的时候判断节点是否存在，如果存在，就不创建，如果不存在，则创建

neo4j里面有一个merge的操作，能够自动帮助我们来判断节点和边在数据库里面属否存在，dgraph有没有呢，答案是有的，就是upsert，我的示例代码：

upsert {
      query {
        var(func: eq(person_id, "xxx")) {
          Person as uid
        }
        var(func: eq(company_id, "yyy")) {
          Company as uid
        }
      }
      mutation {
          set {
           uid(Person) <name> "罗志祥" .
           uid(Person) <person_id> "xxx" .
           uid(Person) <manager> uid(Company) (position="CEO") .
           uid(Person) <dgraph.type> "Person" .
           uid(Company) <name> "无名公司" .
           uid(Company) <dgraph.type> "Company" .
           uid(Company) <company_id> "yyy" .
          }
      }
    }

来看看官网怎么解释的：

To use ​​upsert operations​​​ on a predicate, specify the ​​@upsert​​​ directive in the schema. When committing transactions involving predicates with the ​​@upsert​​ directive, Dgraph checks index keys for conflicts, helping to enforce uniqueness constraints when running concurrent upserts.

我的schema为：

<person_id>: string @index(exact) @upsert .
    <name>: string @index(term) .
    <company_id>: string @index(hash) @upsert .

是不是很简单呢。

怎么验证呢，多moute几次，然后用下面的语句查一下，看有没有重复：

{
   data(func: eq(person_id, "xxx")) {
     name
    person_id
     manager @facets(orderasc:position) {
       name
      company_id
     }
   }
}

参考文献

[1]. ​​https://dgraph.io/docs/master/mutations/#upsert-block​​

[2].External IDs and Upsert Block. ​​https://dgraph.io/docs/mutations/#external-ids-and-upsert-block​​