Though MongoDB offers no restriction on the schema of the documents, we may need some restrictions on it as per our use-cases. MongoDB provides validators for this specific purpose.

There are two validation levels:

1. Strict (default): It means to validate all the documents while inserting and updating.
2. Moderate: It means to validate all the documents while inserting; and to validate the documents that were valid before, during updates.

There are two validation actions that can be performed:

1. Error (default): Rejects the document to be inserted/updated if the validation fails.
2. Warn: Logs the validation error and allows the document to be inserted/updated.

Let's see an example:

Suppose we have an employees database. Let's create an employeeDetails collection with name, email, phone_no and address fields. The validation that needs to be applied is:

• All fields are required.
• name is a string.
• email is a string.
• phone_no is an int.
• address is a string.

Let's apply a validator for the above conditions and create a collection keeping the validationLevel as moderate:

employees> db.createCollection("employeeDetails", {validator: {
$jsonSchema: {
         bsonType: "object",
         required: ["name", "email", "phone_no", "address"],
         properties: {
             name: {
                 bsonType: "string",
                 description: "must be a string and is required."
             },
             email: {
                 bsonType: "string",
                 description: "must be a string and is required."
             },
             phone_no: {
                 bsonType: "int",
                 description:"must be a long and is required."
             },
             address: {
                 bsonType: "string",
                 description:"must be a string and is required."
             }
         }
     }
},
validationLevel: "moderate"})
output> { ok: 1 }

Now, let's insert employee details:

employees> db.employeeDetails.insertOne({
name: "Employee 1", 
email: "employee1@gmail.com", 
phone_no: 1234, 
address: "Address 1"})
output> {
  acknowledged: true,
  insertedId: ObjectId("61bae54e09662dcd9fc9d763")
}

Since these details are valid, they're inserted. Let's try one more time:

employees> db.employeeDetails.insertOne({
name: "Employee 2", 
email: "employee2@gmail.com", 
phone_no: "5678", 
address: "Address 2"})
output> Uncaught:
MongoServerError: Document failed validation
Additional information: {
  failingDocumentId: ObjectId("61bae56b09662dcd9fc9d764"),
  details: {
    operatorName: '$jsonSchema',
    schemaRulesNotSatisfied: [
      {
        operatorName: 'properties',
        propertiesNotSatisfied: [ { propertyName: 'phone_no', details: [ [Object] ] } ]
      }
    ]
  }
}

The validation failed this time because the phone_no provided was string but as per the validator, it should be int. Similarly, we can use different options with validators, validationLevel and validationAction.

---

If you want to add a validator to an existing collection, we can do that with the help of collMod command with the validator option.

Check the official documentation for more details.

Thanks for reading!