การเพิ่มตัวตรวจสอบให้คอลเลกชัน

MongoDB is schema-flexible by default—any document can be inserted regardless of its shape. While this is useful in development, production databases need guardrails to prevent malformed data. MongoDB's schema validation feature lets you attach a JSON Schema rule set to a collection so that only well-formed documents can be inserted or updated, catching data quality problems at the database layer before they propagate.

MongoDB uses the industry-standard JSON Schema specification (draft 4) to express validation rules. You define a $jsonSchema object that declares which fields are required, what types they must be, and any additional constraints. The same format is used everywhere JSON Schema appears—in OpenAPI specs, form libraries, and now MongoDB validators.

Pass a validator option when calling db.createCollection(). The validator contains a $jsonSchema document. The example below requires every users document to have a name (string) and an email (string), and optionally an age (integer).

Use the collMod (collection modification) command to attach or update a validator on a collection that already exists and may already contain data. This does not validate existing documents by default—it only applies to future writes unless you also change validationLevel.

By default, if a document violates the validator, MongoDB rejects the write and throws an error: Document failed validation. The error includes a details field explaining exactly which rule was broken, making it easy to diagnose and fix the offending document. The insert or update is rolled back completely—no partial writes occur.

To inspect the validator attached to a collection, use db.getCollectionInfos() or query the system.js namespace. The returned document includes the full validator specification under options.validator, allowing you to review, copy, or compare validators across environments.

To remove all validation from a collection, run collMod with an empty validator object. This returns the collection to its default schema-free state. You might do this temporarily during a bulk data migration or permanently when retiring validation in favour of application-layer checks.

Mongoose has its own schema validation at the ODM layer that runs in JavaScript before sending data to MongoDB. However, Mongoose validation can be bypassed with insertMany or direct driver calls. Adding a JSON Schema validator at the database level creates an additional safety net that no client can bypass, regardless of the driver or language used.

JSON Schema validators can reach into embedded sub-documents. Use the properties key to define rules for nested fields, and mark the nested object itself with bsonType: 'object'. This allows you to validate every level of a hierarchical document.

To validate that an array field contains only documents of a specific shape, use items inside the property definition. Each element of the array will be validated against the items schema. This is useful for enforcing the shape of embedded line items, tags, or address arrays.

MongoDB Atlas provides a graphical interface for building and editing collection validators without writing raw JSON. Under the collection's Schema tab you can add properties, set types, and mark required fields through a form. Atlas also shows a validation score—the percentage of existing documents that pass the current schema—helping you measure data quality before enforcing strict validation.

Test your understanding of MongoDB & NoSQL Databases concepts from this lesson.

In this lesson you learned: JSON Schema validators are attached via createCollection or collMod, validation failures reject the write and return a detailed error, and nested objects and arrays can also be validated within the same schema document. Next up we explore type, required, and enum constraints in depth.