Question

There seems to be almost no documentation about how to design databases for MongoDB. So I thought I’ll start by posting my questions here.

Assume this collection (fruits_inventory) as an example:

{
    "name"      : "Orange",
    "type"      : "citric",
    "available" : 3
}
{
    "name"      : "Apple",
    "type"      : "pome",
    "available" : 0
    "note"      : "Not shipping this month"
}
{
    "name"      : "Pear",
    "type"      : "pome",
    "available" : 2
}

(No indexes set)

1) Field selection

db.fruits_inventory.findOne({name:"Orange"},{"note":1});

Will this query seek for a document containing only a field name with value Orange and return with the first hit, even if it has no note field set? Or will it keep searching for a document that does contains a note field?

2) With unique indexes

If I set a unique index on name, would the answer for the previous question change?

Only these two questions for now. Answers will be greatly appreciated.

Answer 1

I wrote the following script:

// sofruit.js
db = db.getSiblingDB('test');

db.fruits_inventory.drop();
db.fruits_inventory.save({
    "name"      : "Orange",
    "type"      : "citric",
    "available" : 3
});
db.fruits_inventory.save({
    "name"      : "Apple",
    "type"      : "pome",
    "available" : 0,
    "note"      : "Not shipping this month"
});
db.fruits_inventory.save({
    "name"      : "Pear",
    "type"      : "pome",
    "available" : 2 
});

var a1 = db.fruits_inventory.findOne({name:"Orange"},{"note":1});

db.fruits_inventory.ensureIndex({name:1}, {unique:true});

var a2 = db.fruits_inventory.findOne({name:"Orange"},{"note":1});

Then I ran it from the mongo shell and got:

> load('../mongojs/sofruit.js');
> a1
{ "_id" : ObjectId("4e7d119e9b3e59bf2e0c5199") }
> a2
{ "_id" : ObjectId("4e7d119e9b3e59bf2e0c5199") }  
>

So, the answer is “Yes,” it will return the first hit, even if it has no “note” field set. Adding an index doesn’t change that.