2 
{
"batters":
    {
    "batter":[
            { "id": "1001", "type": "Regular" },
            { "id": "1002", "type": "Chocolate" },
            { "id": "1003", "type": "Blueberry" },
            { "id": "1004", "type": "Devil's Food" }
    ]
    },
    "topping":[
            { "id": "5001", "type": "None" },
            { "id": "5002", "type": "Glazed" },
            { "id": "5005", "type": "Sugar" },
            { "id": "5007", "type": "Powdered Sugar" },
            { "id": "5006", "type": "Chocolate with Sprinkles" },
            { "id": "5003", "type": "Chocolate" },
            { "id": "5004", "type": "Maple" }
     ]
}

基本上要在这里进行全文搜索，我需要对“batters.batter”和“batters.topping”（即两个属性）进行索引。如何处理这种全文搜索。请解释该方法，我将通过 REST API 实现我的搜索。提前谢谢你。

4

1 回答 1

5

解决这个问题的最好方法是稍微改变数据布局，因为全文索引只能对一个属性起作用，并且请求两次索引无论如何都不会很快。因此，我们使用匿名图将字符串连接到它们的对象。

因此，我们创建了两个（顶点）集合，一个边缘集合，一个具有全文索引的顶点集合：

db._create("dishStrings")
db._createEdgeCollection("dishEdges")
db._create("dish")

db.dishStrings.ensureIndex({type: "fulltext", fields: [ "name" ]});

并将文档与将它们联系在一起的关系保存给它们。我们使用_key用于引用_from和_to边关系中的顶点的属性：

db.dishStrings.save({"_key": "1001", "name": "Regular" , type: "Batter"});
db.dishStrings.save({"_key": "1002", "name": "Chocolate", type: "Batter" });
db.dishStrings.save({"_key": "1003", "name": "Blueberry", type: "Batter"});
db.dishStrings.save({"_key": "1004", "name": "Devil's Food", type: "Batter"});
db.dishStrings.save({"_key": "5001", "name": "None", type: "Topping"});
db.dishStrings.save({"_key": "5002", "name": "Glazed", type: "Topping"});
db.dishStrings.save({"_key": "5005", "name": "Sugar", type: "Topping"});
db.dishStrings.save({"_key": "5007", "name": "Powdered Sugar", type: "Topping"});
db.dishStrings.save({"_key": "5006", "name": "Chocolate with Sprinkles", type: "Topping"});
db.dishStrings.save({"_key": "5003", "name": "Chocolate", type: "Topping"});
db.dishStrings.save({"_key": "5004", "name": "Maple", type: "Topping"});

db.dishEdges.save("dishStrings/1001", "dish/batter", {tasty: true, type: "Batter"})
db.dishEdges.save("dishStrings/1002", "dish/batter", {tasty: true, type: "Batter"})
db.dishEdges.save("dishStrings/1003", "dish/batter", {tasty: true, type: "Batter"})
db.dishEdges.save("dishStrings/1004", "dish/batter", {tasty: true, type: "Batter"})
db.dishEdges.save("dishStrings/5001", "dish/batter", {tasty: true, type: "Topping"})
db.dishEdges.save("dishStrings/5002", "dish/batter", {tasty: true, type: "Topping"})
db.dishEdges.save("dishStrings/5003", "dish/batter", {tasty: true, type: "Topping"})
db.dishEdges.save("dishStrings/5004", "dish/batter", {tasty: true, type: "Topping"})
db.dishEdges.save("dishStrings/5005", "dish/batter", {tasty: true, type: "Topping"})
db.dishEdges.save("dishStrings/5006", "dish/batter", {tasty: true, type: "Topping"})
db.dishEdges.save("dishStrings/5007", "dish/batter", {tasty: true, type: "Topping"})

db.dish.save({_key: "batter", tasty: true})

我们重新验证全文索引是否有效：

db._query("FOR oneDishStr IN FULLTEXT(dishStrings, 'name', 'Chocolate')" +
          " RETURN oneDishStr").toArray()

（.toArray()将在控制台上打印我们的结果） 我们得到 3 个安打，1 个击球手，2 个浇头。由于搜索字符串可能包含未经验证的字符串，我们宁愿使用绑定变量来规避注入：

db._query("FOR oneDishStr IN FULLTEXT(dishStrings, 'name', @searchString) " + 
          " RETURN oneDishStr", 
          {searchString: "Chocolate"});

现在让我们使用边缘关系来找到连接的碟子：

db._query("FOR oneDishStr IN FULLTEXT(dishStrings, 'name', @searchString) "+ 
          "RETURN {str: oneDishStr, " + 
                  "dishes: NEIGHBORS(dishStrings, dishEdges, oneDishStr," + 
                                     " 'outbound')}",
           {searchString: "Chocolate"})

这是使用图的旧（最高 2.7）方式，因为我们想使用快速过滤器，让我们将其转换为新的 2.8 语法：

db._query("FOR oneDishStr IN FULLTEXT(dishStrings, 'name', @searchString) " + 
          "  FOR v IN 1..1 OUTBOUND oneDishStr dishEdges RETURN " + 
          "    {str: oneDishStr, dish: v}",
         {searchString: "Chocolate"})

我们可以看到，在这两种情况下，对于Chocolate. 现在我们只对类型为 的命中感兴趣Toppings，因此我们将过滤所有非类型的边Topping：

db._query("FOR oneDishStr IN FULLTEXT(dishStrings, 'name', @searchString) "+
          "   FOR v, e IN 1..1 OUTBOUND oneDishStr dishEdges " + 
          "      FILTER e.type == 'Topping' " +
          "         RETURN {str: oneDishStr, dish: v}", 
          {searchString: "Chocolate"})
于 2016-01-15T13:50:58.670 回答