如何在MongoDB的update调用中覆盖整个JSON文档？$set作用及自定义问询

Hey there! Let's unpack your questions about MongoDB's update function, $set, and document overwrites—this is a common point of confusion, so you're not alone.

What does $set actually do? #

The $set operator is MongoDB's go-to tool for partial document updates. When you use it, only the fields you explicitly specify get updated to their new values, while every other existing field in the document stays exactly as it was. No accidental data loss from fields you didn't mention!

For example, say you have this document:

{
  _id: ObjectId("60d21b4667d0d8992e610c85"),
  username: "jdoe",
  email: "jdoe@olddomain.com",
  age: 30
}

If you run this update:

db.users.updateOne(
  { _id: ObjectId("60d21b4667d0d8992e610c85") },
  { $set: { email: "jdoe@newdomain.com", age: 31 } }
)

The resulting document will still have the username field intact—only email and age get updated. Skip $set and pass fields directly, and you'll replace the entire document (more on that next).

How to overwrite an entire document (keeping only the original _id) #

If you want to replace the whole document except for the _id field, you have two straightforward, reliable options:

Option 1: Use replaceOne() (the most explicit choice) #

replaceOne() is built exactly for this scenario—it replaces the matched document with a new one, automatically preserving the original _id (as long as you don't include a different _id in your new document).

Example:

db.users.replaceOne(
  { _id: ObjectId("60d21b4667d0d8992e610c85") },
  {
    username: "john_doe",
    email: "john@newdomain.com",
    phone: "555-1234", // New field not present in the original
    location: "New York"
  }
)

The original _id stays, and every other field is replaced with the ones in your new document. Any fields from the original that aren't in the new document get removed entirely.

Option 2: Use updateOne() without operators #

You can also use updateOne() and pass the new document directly as the second argument (no $set or other operators). This works the same way as replaceOne() for overwriting, but replaceOne() is more readable for this specific use case.

Example:

db.users.updateOne(
  { _id: ObjectId("60d21b4667d0d8992e610c85") },
  {
    username: "john_doe",
    email: "john@newdomain.com",
    phone: "555-1234"
  }
)

Just make sure you don't include an _id in the new document unless it's identical to the one in your filter—otherwise, MongoDB will throw an error.

Customizing for your use case #

If you need a flexible way to switch between partial updates (with $set) and full document overwrites, you can create a simple helper function to handle both scenarios. Here's a JavaScript example:

async function modifyDocument(collection, filter, data, overwrite = false) {
  if (overwrite) {
    // Remove _id from the data to avoid conflicts with the original document's _id
    const { _id, ...newDocument } = data;
    return await collection.replaceOne(filter, newDocument);
  } else {
    // Use $set to only update the specified fields
    return await collection.updateOne(filter, { $set: data });
  }
}

Now you can call this function with overwrite: true when you want to replace the entire document, or leave it as false (default) for safe partial updates.

Quick recap to avoid mix-ups: #

• $set = partial update (preserves all unspecified fields)
• replaceOne() / direct updateOne() = full overwrite (replaces all fields except _id)
• When overwriting, any fields not included in the new document will be deleted from the existing document.

内容的提问来源于stack exchange，提问作者J.Doe