Hello.
I found that I can't use in find() with id of type ObjectID (find() can accept only integer or string). But in other case if I pass string like 553faa667f09fe5822e595c4 - adapter can't find record and throw error.
So we need to wrap id with ObjectID inside find().
It seems this is an issue.

The library is broken with version mongodb@3 lib. This is a task to upgrade the library's use of mongodb to be compatible with the latest mongodb lib.

This will be a breaking change and major version bump.

What is the right syntax for findAll queries with contains

im trying to query using CONTAINS as filter but it not work's

Sample data structure:

{
    "_id": "1b2c0a744babcc342b27cf9e4c0fb351",
    "id": "1b2c0a744babcc342b27cf9e4c0fb351",
    "type": "class",
    "items": [ 
        "cce40b60d897259857303650c56421adb6508621", 
        "1faeb13afa5ea2feed1b3a5d10b1b1a771238b26", 
        "5a4071b313e67a11f50724449589d0901b9a14bb", 
        "9d083f45fe98a4af4d4aec500d2fe427c6370f91", 
        "5878005cfef2c0f8164a593672e722158405bb4d", 
        "50d2f50e5551651065a3cab188ed68b763c180e3", 
    ],
    "created_at": 1462998180437,
    "update_at": 1477923133426,
},
...

Sample code to run this query:

const JSData = require('js-data');
const DSAdapter = require('js-data-mongodb');

const store = new JSData.DS();
const adapter = new DSAdapter({ uri: 'mongodb://localhost/shop' });
store.registerAdapter('main', adapter, { default: true });

Resource = dataStore.defineResource({ name: 'invoice', idAttribute: '_id', table: 'invoice' });

const params = {
  where: {
    items: {
      contains: '50d2f50e5551651065a3cab188ed68b763c180e3'
    }
  }
}

Resource.findAll(params)
  .then(result => {
    console.log('Result', result.length, result)
  })
  .catch(err => {
    if (err.message.toLowerCase().includes('not found')) return console.log(err);
    throw err;
  });

With the same data structure but with redis adapter and DB work's fine.

I even tried to run some queries with "contains" against strings, with the mongodb sample data - https://docs.mongodb.com/getting-started/shell/import-data/

Thanks, in advance.

Note: using v 0.7.0 (which is weird because in the repo there is only available up to 0.6.0, before passing to the rc 1.0...)

Description

versions:

js-data: 2.10.0
js-data-mongodb: 0.7.0

Steps to reproduce

• Define User resource model
• Assume that a user data:
  { _id: 'XXXX', groups: { groupOne: true, groupTwo: true } }
• Do: User.update('XXXX', {groups: {groupOne: null}})
• Expected result: groupOne gets removed from groups
• End result: groups was overridden and become groups: {groupOne: null}

If there's no way to do atomic update, the code won't scale.

Thanks!

There are cases where someone may implement js-data-mongodb towards a pre existing Mongo database collection and some libraries (ORM's) that perform relational mapping store record _id's in the ObjectID bson value, ie: { "_bsontype": "ObjectID", "id": { "type": "Buffer", "data": [87,...156] } }.

When using js-data-mongodb, it would be convenient to pass an option to translate all ObjectID fields to a string.

I would apply the logic in the method _translateIds() on line 218 and change the name and all references to it to "_translateObjectIDs()".

@jmdobry If you'll accept this feature, I'll put in a PR. Please let me know.

Missing tag for 0.7.0

Please create a tag on commit - af5c2dc0ff1b72a1f2a82ce1d09e9c4f7d773d28 for tag 0.7.0

Assuming that 0.7.0 was the last version that work with js-data v2, and its a version published on npm.

Thanks!

// Why couldn't Mongo just use "id"?

No shit?!

@jmdobry How do you feel about making default functionality to convert id to _id and vica versa with adapter interactions?

We can still have the config idAttribute property to be specified for override.

Description

versions :
js-data: 3.0.0-rc.6
js-data-mongodb : 1.0.0-rc.1

Steps to reproduce

Here is my model schema :

{
  "type": "object",
  "properties": {
    "a": {
      "type": "string"
    },
    "b": {
      "type": "string"
    }
  },
  "required": ["a"],
  "additionalProperties": false
}

When i insert or update not all the properties of my model (example with only property "a" : {a: "a"}), the others properties are set to undefined by js-data. So the input props of methods create or update looks like :

{
  a: "a",
  b: undefined
}

The problem is that the Node.js mongodb driver, insert null in the database for all the undefined properties. So when you just want to update property "a", the previous value of property "b" is erase by a null value.

The workaround solution i found, is to remove all undefined properties of the object in methods beforeCreate and beforeUpdate on the adapter. But i think it could be made directly on insert or update methods.

// Workaround to avoid undefined values to be set to null in MongoDB
const adapter = new MongoDBAdapter({
  uri: config.MONGODB_URI,
  beforeUpdate(mapper, id, props) {
    const removeUndefined = (obj) => {
      Object.keys(obj).forEach(key =>
        (obj[key] && typeof obj[key] === 'object') && removeUndefined(obj[key]) ||
        (obj[key] === undefined) && delete obj[key]
      );
      return obj;
    };
    removeUndefined(props);
  }
});

Solution can be found here : http://stackoverflow.com/questions/286141/remove-blank-attributes-from-an-object-in-javascript. It should be recursive and only remove undefined properties.

Any thoughts ?

Thanks!

See js-data/js-data#110

"hasOne" relation inconsistent return

hasOne relationship is a one to one relation, when a related item exists the OBJECT is returned, but if the related item does not exists on the related collection and empty ARRAY is returned.

I think the query should return an empty object or better yet undefined.

js-data: 2.10
js-data-mongo: 0.7.0

Thanks!

This work fine:

User.findAll({
  where: {
    field: {
      '==': '393fe308fc3d796b6b80c548ceeac69008a68a09'
    }
  }
})
...
User.findAll({
    field: {
      '==': '393fe308fc3d796b6b80c548ceeac69008a68a09'
    }
})
...
User.findAll({
    field: '393fe308fc3d796b6b80c548ceeac69008a68a09'
})

But this not:

User.findAll({
  where: {
    field: '393fe308fc3d796b6b80c548ceeac69008a68a09'
  }
})