[skratchdot.]


MongoDB - schema.js

Project Page
Source Code
Issues

Description:

This is a schema analysis tool for MongoDB. It accomplishes this by extending the mongo shell, and providing a new function called schema() with the following signature:

DBCollection.prototype.schema = function (optionsOrOutString)

The schema() function accepts all the same parameters that the mapReduce() function does. It adds/modifies the following 4 parameters that can be used as well:

wildcards - array (default: [])
    By using the $, you can combine report results.
    For instance: '$' will group all top level keys and
    'foo.$.bar' will combine 'foo.baz.bar' and 'foo.bar.bar'

arraysAreWildcards - boolean (default: true)
    When true, 'foo.0.bar' and 'foo.1.bar' will be
    combined into 'foo.$.bar'
    When false, all array keys will be reported

fields - object (default: {})
    Similar to the usage in find(). You can pick the
    fields to include or exclude. Currently, you cannot
    pass in nested structures, you need to pass in dot notation keys.

limit - number (default: 50)
    Behaves the same as the limit in mapReduce(), but defaults to 50.
    You can pass in 0 or -1 to process all documents.

Usage:

// Return schema results inline
db.users.schema();

// Create and store schema results in the 'users_schema' collection
db.users.schema('users_schema'); // Option 1
db.users.schema({ out: 'users_schema' }); // Option 2
db.users.schema({ out: { replace: 'users_schema' } }); // Option 3

// Only report on the key: 'name.first'
db.users.schema({ fields: { 'name.first': 1 } });

// Report on everything except 'name.first'
db.users.schema({ fields: { 'name.first': -1 } });

// Combine the 'name.first' and 'name.last' keys into 'name.$'
db.users.schema({ wildcards: ['name.$'] });

// Don't treat arrays as a wildcard
db.users.schema({ arraysAreWildcards: false });

// Process 50 documents
db.users.schema();

// Process all documents
db.users.schema({ limit: -1 });

Example Result Set:

> // Start fresh with a new collection called 'users'
> db.users.remove();
>
> // Add a few records with different schemas
> db.users.insert({'name' : {'first' : 'John', 'last' : 'Smith'}, 'isRegistered' : false, 'tags' : ['male']});
> db.users.insert({'name' : {'first' : 'Bob', 'last' : 'Smith'}, 'isRegistered' : false, 'tags' : ['male','new']});
> db.users.insert({'name' : {'first' : 'Amy', 'last' : 'Smart'}, 'isRegistered' : 1, 'tags' : ['female']});
> db.users.insert({'name' : 'Bob Smith', 'isRegistered' : '0', 'tags' : ['male']});
>
> // Print our results to the console
> db.users.schema();
Processing 4 document(s)...
{
  "results" : [
    {
      "_id" : "_id",
      "value" : {
        "wildcard" : false,
        "types" : ["objectid"],
        "results" : [
          {"type" : "all","docs" : 4,"coverage" : 100,"perDoc" : 1},
          {"type" : "objectid","docs" : 4,"coverage" : 100,"perDoc" : 1}
        ]
      }
    },
    {
      "_id" : "isRegistered",
      "value" : {
        "wildcard" : false,
        "types" : ["boolean","number","string"],
        "results" : [
          {"type" : "all","docs" : 4,"coverage" : 100,"perDoc" : 1},
          {"type" : "boolean","docs" : 2,"coverage" : 50,"perDoc" : 1},
          {"type" : "number","docs" : 1,"coverage" : 25,"perDoc" : 1},
          {"type" : "string","docs" : 1,"coverage" : 25,"perDoc" : 1}
        ]
      }
    },
    {
      "_id" : "name",
      "value" : {
        "wildcard" : false,
        "types" : ["bson","string"],
        "results" : [
          {"type" : "all","docs" : 4,"coverage" : 100,"perDoc" : 1},
          {"type" : "bson","docs" : 3,"coverage" : 75,"perDoc" : 1},
          {"type" : "string","docs" : 1,"coverage" : 25,"perDoc" : 1}
        ]
      }
    },
    {
      "_id" : "name.first",
      "value" : {
        "wildcard" : false,
        "types" : ["string"],
        "results" : [
          {"type" : "all","docs" : 3,"coverage" : 75,"perDoc" : 1},
          {"type" : "string","docs" : 3,"coverage" : 75,"perDoc" : 1}
        ]
      }
    },
    {
      "_id" : "name.last",
      "value" : {
        "wildcard" : false,
        "types" : ["string"],
        "results" : [
          {"type" : "all","docs" : 3,"coverage" : 75,"perDoc" : 1},
          {"type" : "string","docs" : 3,"coverage" : 75,"perDoc" : 1}
        ]
      }
    },
    {
      "_id" : "tags",
      "value" : {
        "wildcard" : false,
        "types" : ["array"],
        "results" : [
          {"type" : "all","docs" : 4,"coverage" : 100,"perDoc" : 1},
          {"type" : "array","docs" : 4,"coverage" : 100,"perDoc" : 1}
        ]
      }
    },
    {
      "_id" : "tags.$",
      "value" : {
        "wildcard" : true,
        "types" : ["string"],
        "results" : [
          {"type" : "all","docs" : 4,"coverage" : 100,"perDoc" : 1.25},
          {"type" : "string","docs" : 4,"coverage" : 100,"perDoc" : 1.25}
        ]
      }
    }
  ],
  "timeMillis" : 16,
  "counts" : {"input" : 4,"emit" : 26,"reduce" : 7,"output" : 7},
  "ok" : 1,
}

Caveats:

By design, schema() returns 'bson' rather than 'object'. It will return 'numberlong' rather than 'number', etc.

Installation:

Download: schema.js

Option 1

Add this script to your .mongorc.js file.

See: http://www.mongodb.org/display/DOCS/Overview+-+The+MongoDB+Interactive+Shell#Overview-TheMongoDBInteractiveShell-.mongorc.js

Option 2

Start the shell after executing this script

mongo --shell schema.js

Inspired by:

Variety: https://github.com/JamesCropcho/variety