使用 $set 和 $unset 的 updateOne 与 updateMany

MongoDB provides two primary methods for modifying existing documents:

• updateOne(filter, update, options) — modifies the first document matching the filter. If multiple documents match, only one is updated.
• updateMany(filter, update, options) — modifies all documents matching the filter.

Both methods return a result object with matchedCount (how many documents matched the filter) and modifiedCount (how many were actually changed—some matching documents may already have the target value). Always verify modifiedCount in your application logic.

The $set operator sets the value of one or more fields. If a field does not exist, $set creates it. If it already exists, $set replaces its value. Only the specified fields are changed—all other fields in the document remain untouched.

This is the most important update operator and the one you will use in the vast majority of update operations. Never pass a plain document as the update argument without an operator—doing so replaces the entire document, losing all other fields.

A critical MongoDB beginner mistake: if you pass a plain document as the update argument—without any update operators—MongoDB treats it as a full document replacement. The matched document is replaced entirely with the new document (preserving only _id). All other fields are lost.

This behavior is actually intentional and useful for replacing an entire document, but it is a common source of accidental data loss. Always use update operators ($set, $inc, etc.) unless you explicitly intend to replace the full document.

The $unset operator removes one or more fields from a document. The value you assign in $unset is ignored—conventionally set to '' or 1. The field is completely removed from the document, not set to null.

Use $unset when you want to clean up legacy fields during a schema migration, remove sensitive data (like a password reset token after it is used), or strip fields that should no longer be stored on certain documents.

You can combine multiple update operators in a single updateOne or updateMany call. They are applied atomically in a single operation—there is no window where the document is in a partially updated state.

This is important for schema migrations where you simultaneously add a new field and remove an old one. Doing it in one atomic operation prevents any client from reading a document in the transitional state where both old and new fields exist.

The $rename operator atomically renames a field within a document. It is equivalent to $seting the new name with the old value and then $unseting the old name—but done in a single atomic step.

Use $rename when refactoring your schema and changing field names. It works correctly even when the destination field does not exist. If the source field does not exist, $rename does nothing (no error). It cannot rename fields across embedded documents.

The $currentDate operator sets a field to the current date/time at the moment the update is executed on the server. Using server-side timestamps is more reliable than passing a date from the application, because it avoids clock skew between multiple application servers.

Set the value to true to use the BSON Date type, or { $type: 'timestamp' } for the BSON Timestamp type (used internally by the replication system). In application code, BSON Date is almost always what you want.

Setting { upsert: true } in the options tells MongoDB: if no document matches the filter, insert a new document that combines the filter fields and the update. This is the upsert (update-or-insert) pattern—useful for 'insert if not exists, update if exists' semantics without a separate read operation.

The inserted document contains the fields from both the filter and the $set. Check result.upsertedId to see if an insert occurred. Upserts are atomic—no race condition between checking existence and inserting.

updateMany applies the same update to all matching documents in one operation. This is how you perform bulk schema migrations, add default values to missing fields, or update a batch of records based on a condition.

Be careful with updateMany({}, ...) (empty filter)—it modifies every document in the collection. Always include a filter that limits the scope. Consider adding a condition like { newField: { $exists: false } } to only process documents that need the change.

findOneAndUpdate(filter, update, options) atomically finds a document, updates it, and returns it in a single operation. This eliminates the race condition that exists when you do a separate findOne followed by an updateOne.

By default it returns the document before the update. Set { returnDocument: 'after' } to get the updated version. Combine with upsert: true to find-or-create patterns. This is ideal for claiming queue items, implementing counters, and any pattern where you need to both modify and read the document atomically.

MongoDB provides many update operators beyond and . Here is a quick reference of the most important ones you will encounter:

• : Set field value (create if absent)
• : Remove a field completely
• : Rename a field atomically
• : Increment/decrement a numeric field
• : Multiply a numeric field
• /: Conditionally update to smaller/larger value
• /: Add/remove array elements
• : Add to array only if unique
• : Set to server-side current date/time

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

In this lesson you learned: $set modifies specific fields without touching others—always use it rather than passing a bare document to avoid accidental full-document replacement, $unset removes fields entirely and is essential for schema migrations and cleaning up sensitive temporary data like password reset tokens, and upsert: true combined with $setOnInsert enables atomic find-or-create patterns without race conditions. Next up we explore numeric mutation operators: $inc, $mul, $min, and $max.