MongoDB找到内部数组

Question

我是MongoDB的新手，我想过滤一个内部数组。当前文档的样子：

{ 
    "data1": "value1", 
    "data2": "value2", 
    ... 
    "array1": [ 
     { 
      "field1": field1value" 
      "field2": field2value" 
     }, 
     { 
      "field1": expectedvalue" 
      "field2": field2value" 
     }, 
    ] 
} 

我期待像一个结果：

{ 
    "data1": "value1", 
    "data2": "value2", 
    ... 
    "array1": [ 
     { 
      "field1": expectedvalue" 
      "field2": field2value" 
     }, 
    ] 
} 

我试图找到并聚集在In MongoDB, how do I search in an array of sub-documents?和其他类似的问题建议。

的问题是，我失去所有的信息的阵列（数据1，数据2，...）以外：

{ 
    "_id" : { "$oid" : "8001212651b8a68278edbc92"}, 
    "array1": [ 
     { 
      "field1": expectedvalue" 
      "field2": field2value" 
     }, 
    ] 
} 

Answer 1

要仅在array1阵列相匹配，你必须给定的条件，其在返回的子文档project和filter（可用于版本> = 3.2）。这里有一个例子：

db.collection.aggregate([ 
    // this is optional, it is only included here to show you that you _can_ match 
    // the documents before you match the array of sub documents 
    {$match: {data1: 'value1'}}, 
    {$project: { 
      data1: 1, 
      data2: 1, 
      array1: {$filter: { 
       input: '$array1', 
       as: 'a', 
       cond: {$eq: ['$$a.field1', 'expectedvalue']} 
      }} 
     }} 
]) 

这将返回如下：

{ 
    "_id" : ..., 
    "data1" : "value1", 
    "data2" : "value2", 
    "array1" : [ 
     { 
      "field1" : "expectedvalue", 
      "field2" : "field2value" 
     } 
    ] 
} 

您链接到答案这样的情况下，你需要通过Java驱动程序来运行此查询涉及蒙戈Java驱动程序，在这里它是：

List<Document> documents = collection.aggregate(Arrays.asList(
      new Document("$project", new Document() 
        // include data1 
        .append("data1", 1) 
        // include data2 
        .append("data2", 1) 
        // include only those elements of array1 which match the filter condition 
        .append("array1", new Document("$filter", 
          new Document("input", "$array1") 
            .append("as", "a") 
            .append("cond", new Document("$eq", Arrays.asList("$$a.field1", "expectedvalue")))))) 
    )).into(new ArrayList<>());