Best way to perform a full text search in MongoDB

问题:

I'm searching on Google since days and I tried many things but I still can not perform a good full text search on my user collection.

I tried ElasticSearch but was pretty impossible to query and paginate...

I tried many plugins for Mongoose like ElMongo, mongoose-full-text, Mongoosastic, etc... everyone are really bad documented and I don't know how to perform a good full text search.

So, my collection is a normal collection:

user = {
  name: String,
  email: String,
  profile: {
    something: String,
    somethingElse: String
  }
}

I have a search input in a page with a simple POST, if I type hello world what I need is to search on the entire collection fields the matching words of my search query and get the results.

It will be really nice also to have options to handle a pagination like 10 items per page or something...

What is the best solution to achieve this? I'm using MongoDB 2.6.* with Mongoose, NodeJS and ExpressJS.

Thanks.

回答1:

You can add a text index to your Mongoose schema definition that lets you use the $text operator in your find queries to search all fields included in the text index.

To create an index to support text search on, say, name and profile.something:

var schema = new Schema({
  name: String,
  email: String,
  profile: {
    something: String,
    somethingElse: String
  }
});
schema.index({name: 'text', 'profile.something': 'text'});

Or if you want to include all string fields in the index, use the '$**' wildcard:

schema.index({'$**': 'text'});

This would enable you to performed a paged text search query like:

MyModel.find({$text: {$search: searchString}})
       .skip(20)
       .limit(10)
       .exec(function(err, docs) { ... });

For more details, read the full MongoDB Text Indexes documentation.

回答2:

For search query you can use Elasticsearch plugin. In Elasticsearch you can make your own custom search query for searching any particular text and string with the specified time duration in the real time.

But before that you have to make indexing of your json docs residing in MongoDB. You have to connect MongoDB with Elasticsearch server.then using the river functionality of Elasticsearch you have to make the indexing in Elasticsearch server of your data, then only you can make your custom query for searching.