I'm using this schema with mongoose 3.0.3
from npm:
var schema = new Schema({
_id: Schema.ObjectId,
email: {type: String, required: true, unique: true}
});
If I try to save a email that is already in db, I expect to get a ValidationError
like if a required
field is omitted. However this is not the case, I get a MongoError: E11000 duplicate key error index
.
Which is not a validation error (happens even if I remove the unique:true).
Any idea why?
This is expected behavior
The unique: true
is equivalent to setting an index in mongodb like this:
db.myCollection.ensureIndex( { "email": 1 }, { unique: true } )
To do this type of validation using Mongoose (Mongoose calls this complex validation- ie- you are not just asserting the value is a number for example), you will need to wire in to the pre-save event:
mySchema.pre("save",function(next, done) {
var self = this;
mongoose.models["User"].findOne({email : self.email},function(err, results) {
if(err) {
done(err);
} else if(results) { //there was a result found, so the email address exists
self.invalidate("email","email must be unique");
done(new Error("email must be unique"));
} else {
done();
}
});
next();
});
I prefer putting it in path validation mechanisms, like
UserSchema.path('email').validate(function(value, done) {
this.model('User').count({ email: value }, function(err, count) {
if (err) {
return done(err);
}
// If `count` is greater than zero, "invalidate"
done(!count);
});
}, 'Email already exists');
Then it'll just get wrapped into ValidationError
and will return as first argument when you call validate
or save
.