Reputation: 769
I originally have these two schemas:
var UserSchema = new Schema({
first: String,
last: String
});
var SaleSchema = new Schema({
createdAt: Date,
registeredBy: { type: Schema.ObjectId, ref: 'User' }
});
But I want to edit my SaleSchema to save the user name instead of the ID, so I changed it for:
var SaleSchema = new Schema({
createdAt: Date,
registeredBy: String
});
Next, I wanted to edit all the Sales documents and replace the user IDs on registeredBy for the user's full name, but I can't seem to be able to perform a find
query for the old ID's.
Long story short, this query returns no matches on mongoose, but it works perfectly using the mongo console:
Mongoose
Sale.find({ registeredBy: '57ea0cbb47431f0b43b87d42' })
.then(results => res.json(results))
.catch(err => res.status(500).json(err));
// result: []
MongoDB console
db.sales.find({ registeredBy: '57ea0cbb47431f0b43b87d42' })
// result: 8 elements
After I modify my schema's property back to ObjectId, the mongoose query works again. Since I need to migrate to a new datatype, I want to be able to query and store both types of values. Is this possible?
Upvotes: 0
Views: 450
Reputation: 79208
Good question this is a complicated edge case. I am not super familiar with Mongoose specifically, but one way to do this would be to migrate your data at a lower level. For example, create a mongo script that uses the low-level mongo API to do the migration, something along the lines of:
db.sales.find().forEach(function(doc){
var user = db.users.find({ _id: doc.registeredBy });
db.sales.update({ _id: doc._id, }, {
$set: { registeredBy: user.first + ' ' + user.last }
});
});
This is similar to what a module like https://github.com/balmasi/migrate-mongoose does, but I've personally found it easier to use mongo scripts on the cli directly.
mongo < sale-schema-migration.js
Upvotes: 1