MongoDB writing a query for search engine

Question

I am trying to write a search script in MongoDB but can't figure out how to do it....The thing I wan't to do is as follows....

Lets I have a string array XD = {"the","new","world"}

Now i want to search string array XD in MongoDB document (using regex) and get the result document. For example..

{ _id: 1, _content: "there was a boy" }
{ _id: 2, _content: "there was a boy in a new world" }
{ _id: 3, _content: "a boy" }
{ _id: 4, _content: "there was a boy in world" }

now I want to get result in accordance to _content contains the string in string array XD

{ _id: 2, _content: "there was a boy in a new world", _times: 3 }
{ _id: 4, _content: "there was a boy in world", times: 2 }
{ _id: 1, _content: "there was a boy", times: 1 }

as first document (_id : 2 ) contains all three { "the" in there, "new" as new, "world" as world } so it got 3

second document (_id: 4) only two { "world" as world } so it got 2

Have you looked at text searching? Typically words like "the" are ignored (and arguably should be) but "new" and "world" would match and rank just like you are expecting already. — Blakes Seven
– Blakes Seven, Commented Feb 9, 2016 at 5:43
its just an example......word can be anything.....and here i am trying to use regex ( like if i want to search "exam" than document containing "example" or "exammed" should return....here i am talking about aggregate function which could return these type of document — Shiid Shah Sury
– Shiid Shah Sury, Commented Feb 9, 2016 at 5:51
Just suggesting that maybe elasticsearch would be more suitable for your needs. Related question — Mario Trucco
– Mario Trucco, Commented Feb 9, 2016 at 6:34
thanks.....but here i also wanted to know that which document contains how many of string in string array, so that i can ranked them. — Shiid Shah Sury
– Shiid Shah Sury, Commented Feb 9, 2016 at 6:41

dikesh · Accepted Answer · 2016-02-09 08:36:51Z

Here is what you can do.

Create a Regex to be matched with _content

XD = ["the","new","world"];
regex = new RegExp(XD.join("|"), "g");

Store a JS function on the server, which matches the _content with XD and returns the counts matched

db.system.js.save(
   {
     _id: "findMatchCount",
     value : function(str, regexStr) {
        XD = ["the","new","world"];
        var matches = str.match(regexStr);
        return (matches !== null) ? matches.length : 0;
     }
   }
)

Use the function with mapReduce

db.test.mapReduce(
    function(regex) {
       emit(this._id, findMatchCount(this._content, regex));
    },
    function(key,values) {
        return values;
    },
    { "out": { "inline": 0 } }
);

This will produce the output as below:

{
    "results" : [
        {
            "_id" : 1,
            "value" : 1
        },
        {
            "_id" : 2,
            "value" : 1
        },
        {
            "_id" : 3,
            "value" : 1
        },
        {
            "_id" : 4,
            "value" : 1
        }
    ],
    "timeMillis" : 1,
    "counts" : {
        "input" : 4,
        "emit" : 4,
        "reduce" : 0,
        "output" : 4
    },
    "ok" : 1
}

I am not sure how efficient this solution is but it works.

Hope this helps.

Collectives™ on Stack Overflow

MongoDB writing a query for search engine

1 Answer 1

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related