Why should you care?

YouTube-like apps need fast metadata queries, resume-watching, and recommendations — not BLOBs inside Mongo documents.

See it live — copy this example

Open mongosh or MongoDB Compass, select database nosqlverse, then run the example. Change one field and run again.

db.videos.insertOne({
  _id: "vid-100",
  title: "MongoDB Sharding Explained",
  channelId: "ch-1",
  durationSec: 842,
  tags: ["mongodb", "sharding"],
  s3Key: "videos/vid-100/master.m3u8",
  publishedAt: new Date()
})
db.watchProgress.updateOne(
  { userId: "u1", videoId: "vid-100" },
  { $set: { positionSec: 120, updatedAt: new Date() } },
  { upsert: true }
)
db.videos.createIndex({ channelId: 1, publishedAt: -1 })
db.videos.find({ channelId: "ch-1" }).sort({ publishedAt: -1 }).limit(20)
db.watchProgress.findOne({ userId: "u1", videoId: "vid-100" })

Run Example »

Edit the code below and click Run to see the result in Toolliyo’s live editor.

What happened?

• videos store metadata and an S3 key to the stream.
• watchProgress upserts resume position per user/video.
• Channel pages query by channelId sorted by publishedAt.
• Media bytes stay out of Mongo.

Practice next

1. Insert video metadata pointing at a fake s3Key.
2. Upsert watch progress twice and confirm overwrite.
3. List a channel’s videos newest first.
4. Add unique index { userId: 1, videoId: 1 } on watchProgress.
5. Record a views counter with $inc on videos.

Remember

Mongo = metadata + progress; object store = media. Index channel timelines. Upsert resume positions.

Course video player

NoSQLVerse lessons resume at positionSec when the learner returns.

Outcome: Completion rates rise because progress is never lost.