Question

I have the following layout:

{
    "URL": "http://someurl.de",
    "plugins": {
        "HTTPServer": {
            "os": [
                "FreeBSD"
            ],
            "string": [
                "Apache/2.2.21 (FreeBSD) mod_ssl/2.2.21 OpenSSL/0.9.8q DAV/2 PHP/5.3.8 with Suhosin-Patch"
            ]
        }
    }
}

From which I want to get the count of the unique items stored in plugins.HTTPServer.string. All the MapReduce examples however just refer to single level documents. As I understand the examples, you have to emit the data in the map function (or choose data you want to extract) and then reduce is used to further process the results. I think my problem is during the mapping stage – that I need to get access to the string value above that reads: “Apache/2.2…”

As I’ve only spent the last day in MongoDB, forgive my ignorance if I’m not asking the right questions here. Am I headed in the right direction? I know I can use distinct = db.coll.distinct(‘plugins.HTTPServer.string’), but I’d like to get this done with MapReduce.

map = function() {
  server = this.plugins.HTTPServer.string
  emit({server : this.server}, {count: 1});
}

reduce = "function(key, values) {
  var count = 0;

  values.forEach(function(v) {
    count += v['count'];
  });

  return {count: count};
}"

Answer 1

You’ve got a few problems:

1. this.server in the map funtion’s emit should just be server
2. In your documents, the "string" field is an array, not a single string, so you’re emitting the array as your key which probably isn’t what you want.
3. You’ve got stray " characters in your reduce function.

Try this instead:

var map = function() {
  if (this.plugins && this.plugins.HTTPServer && this.plugins.HTTPServer.string) {
    this.plugins.HTTPServer.string.forEach(function(server) {
      emit({server: server}, {count: 1});
    });
  }
}

var reduce = function(key, values) {
  var count = 0;

  values.forEach(function(v) {
    count += v['count'];
  });

  return {count: count};
}