无共同实体时，如何将多MongoDB集合数据合并为单集合或视图？

Hey there! Let's figure out how to combine your user, posts, and uploads collections into a single view or collection so your frontend can query just one place for all post-related data.

First, let's address the core issue you mentioned: you thought joins weren't possible because there's no common entity, but chances are you just need to leverage MongoDB's $lookup aggregation stage properly—even across multiple collections. Let's break down the solutions:

1. Create a Materialized View or Virtual View with Aggregation #

MongoDB's aggregation pipeline lets you join multiple collections and shape the output exactly like your sample data. Here's how to do it, assuming your collections have logical relationships (which they should, even if you didn't explicitly define them):

• posts should have a userId field linking to the _id of the corresponding user document
• uploads should have a postId field linking to the _id of the post it belongs to

Example Aggregation Pipeline to Build a Combined Collection #

Run this pipeline to generate a new combinedPosts collection with all your required data:

db.posts.aggregate([
  // Join with the user collection to get user details
  {
    $lookup: {
      from: "user",
      localField: "userId",
      foreignField: "_id",
      as: "userDetails"
    }
  },
  // Unwind the user array (since $lookup returns an array, even for single matches)
  { $unwind: "$userDetails" },
  // Join with the uploads collection to get image info
  {
    $lookup: {
      from: "uploads",
      localField: "_id",
      foreignField: "postId",
      as: "uploadDetails"
    }
  },
  // Add calculated fields like "hasImage" and extract image metadata
  {
    $addFields: {
      userName: "$userDetails.userName",
      userRating: "$userDetails.userRating",
      postCount: "$userDetails.postCount",
      hasImage: { $gt: [{ $size: "$uploadDetails" }, 0] },
      imageSize: {
        $cond: {
          if: { $gt: [{ $size: "$uploadDetails" }, 0] },
          then: { $first: "$uploadDetails.imageSize" },
          else: null
        }
      },
      uploadDate: {
        $cond: {
          if: { $gt: [{ $size: "$uploadDetails" }, 0] },
          then: { $first: "$uploadDetails.uploadDate" },
          else: null
        }
      }
    }
  },
  // Keep only the fields your frontend needs
  {
    $project: {
      userName: 1,
      userRating: 1,
      postText: 1,
      postCount: 1,
      hasImage: 1,
      imageSize: 1,
      uploadDate: 1,
      _id: 1 // Optional: keep the post ID if needed
    }
  },
  // Write the result to a new collection (this creates a materialized view)
  { $out: "combinedPosts" }
])

Create a Virtual View (Real-Time) #

If you don't want to maintain a separate collection and prefer real-time data, create a virtual view instead. This view will dynamically run the aggregation every time you query it:

db.createView(
  "combinedPostsView",
  "posts",
  [
    { $lookup: { from: "user", localField: "userId", foreignField: "_id", as: "userDetails" } },
    { $unwind: "$userDetails" },
    { $lookup: { from: "uploads", localField: "_id", foreignField: "postId", as: "uploadDetails" } },
    {
      $addFields: {
        userName: "$userDetails.userName",
        userRating: "$userDetails.userRating",
        postCount: "$userDetails.postCount",
        hasImage: { $gt: [{ $size: "$uploadDetails" }, 0] },
        imageSize: { $cond: { if: { $gt: [{ $size: "$uploadDetails" }, 0] }, then: { $first: "$uploadDetails.imageSize" }, else: null } },
        uploadDate: { $cond: { if: { $gt: [{ $size: "$uploadDetails" }, 0] }, then: { $first: "$uploadDetails.uploadDate" }, else: null } }
      }
    },
    { $project: { userName: 1, userRating: 1, postText: 1, postCount: 1, hasImage: 1, imageSize: 1, uploadDate: 1, _id: 1 } }
  ]
)

2. If There Are No Logical Relationships (Fix Your Data Model) #

If your collections truly have no way to link posts to users or uploads, you'll need to adjust your data model first:

• Add a userId field to every posts document that references the corresponding user's _id
• Add a postId field to every uploads document that references the post it belongs to

MongoDB is a non-relational database, but that doesn't mean you can't have relationships—you just need to define them explicitly for joins to work.

3. Automate Updates with Triggers (For Fresh Data) #

If you use MongoDB Atlas, you can set up database triggers that automatically update the combinedPosts collection whenever a document is added/updated in user, posts, or uploads. This ensures your combined collection always has the latest data without manual pipeline runs.

For self-hosted MongoDB, you can write a simple script (using Node.js/Python/MongoDB Shell) that runs on a schedule to refresh the combined collection.

内容的提问来源于stack exchange，提问作者kbang