mongoDB Map / Reduce

时间:2012-05-01 21:38:25

标签: mongodb mapreduce mongodb-query nosql

我想在mongoDB中找出这个map / reduce系统。我的收藏中有以下基本架构/布局。

{
    _id: 1,
    name: n1,
    website: w1,
    tags: [
        myTag1,
        myTag3
    ]
}

{
    _id: 2,
    name: n2,
    website: w2,
    tags: [
        myTag2,
        myTag3
    ]
}

{
    _id: 3,
    name: n3,
    website: w3,
    tags: [
        myTag2,
        myTag4
    ]
}

如何检索唯一标记数组?我希望将其退还给我以供进一步使用。

{
    tags: [
        myTag1,
        myTag2,
        myTag3,
        myTag4
    ]
}

顺便说一下,这就是我想出来的,但它只返回每个项目的_id和标签,而不是将标签组合成一个对象。

var map = function() {emit( this._id,{tags: this.tags});};

var reduce = function(key, values) {
    var t = [];

    values.forEach(function(doc) {
        var tags = doc.tags;
        tags.forEach(function(tag) {
            if (!(tag in t)) {
                t.push(tag);
            }
        });
    });

    return {tags: t};
};

var op = db.businesses.mapReduce(map, reduce, {out: "mr_results"});

db[op.result].find();

2 个答案:

答案 0 :(得分:3)

在您的情况下,无需使用map-reduce。只需使用distinct功能:

db.businesses.distinct('tags')

您可以在mongo shell中尝试:

> use test
switched to db test
> db.businesses.insert({tags: ['tag1', 'tag2']})
> db.businesses.insert({tags: ['tag3', 'tag4']})
> db.businesses.find()
{ "_id" : ObjectId("4fa05b2b036495bf4ac9c0cc"), "tags" : [ "tag1", "tag2" ] }
{ "_id" : ObjectId("4fa05b33036495bf4ac9c0cd"), "tags" : [ "tag3", "tag4" ] }

> db.businesses.distinct('tags')
[ "tag1", "tag2", "tag3", "tag4" ]

另外,您应该记住,MongoDB中的map / reduce不适合实时查询。

答案 1 :(得分:1)

使用MongoDB MapReduce,您可以按照以下方式执行此操作:

function m() {
  this.tags.forEach(function(x) { emit('tag', x); });
}
function r(k, v) {
  var res = {};
  v.forEach(function(x) { res[x] = true; });
  return res;
}
db.businesses.mapReduce(m, r, {out:'out'});
// Now the 'out' collection has a record whose "value" property
// has a key for each of the tags in the source collection.
function getTags(coll) {
  var tags=[], o=db[coll].find()[0].value;
  for (var i in o) { tags.push(i) }
  return tags; // Same as Object.keys(db[coll].find()[0].value)
}
listTags('out'); // => ['myTag1', 'myTag2', 'myTag3', 'myTag4']