如何使用Map/Reduce从Couchbase获取各ID的最新记录？

Great question! While N1QL is straightforward for this use case, Views are still a solid lightweight, index-based alternative. Let’s walk through exactly how to build this with Map/Reduce:

1. Map Function #

First, we need to structure our emitted data to group records by your target ID and include a sortable timestamp (or date field) to identify the latest entry.

Assuming your documents look like this (adjust field names to match your actual data):

{
  "id": 1000,
  "timestamp": 1698765432, // Unix timestamp (ISO strings like "2023-10-31T12:34:56Z" work too)
  "content": "your record data here"
}

Here’s the Map function to use:

function (doc, meta) {
  // Only process documents that have both the ID and timestamp fields
  if (doc.id && doc.timestamp) {
    // Emit the ID as the grouping key, and an object with timestamp + full document as value
    emit(doc.id, { ts: doc.timestamp, record: doc });
  }
}

This ensures all records are grouped under their respective IDs, and each entry carries the info we need to pick the latest one.

2. Reduce Function #

Next, we need a custom Reduce function that, for each group of records with the same ID, selects the entry with the highest timestamp.

Couchbase’s Reduce functions handle both initial reduction and rereduction (combining intermediate results), so we’ll build support for both cases:

function (keys, values, rereduce) {
  let latestEntry = values[0];
  
  if (rereduce) {
    // Comparing already-reduced results (each is the latest from a subset of the group)
    for (let i = 1; i < values.length; i++) {
      if (values[i].ts > latestEntry.ts) {
        latestEntry = values[i];
      }
    }
  } else {
    // Initial reduction: check all entries in the ID group
    for (let i = 1; i < values.length; i++) {
      if (values[i].ts > latestEntry.ts) {
        latestEntry = values[i];
      }
    }
  }
  
  return latestEntry;
}

This function iterates through the group, tracks the entry with the highest timestamp, and returns that as the latest record for the ID.

3. Querying the View #

Once your View is created, query it with these settings to get your desired results:

• Set group=true: This tells Couchbase to group results by the emitted ID key, so you get one result per unique ID.
• Optional: Add keys=[1000,2000] if you only want results for specific IDs (instead of all IDs in the bucket).

The returned data will be the latest record object for each ID you target.

Quick Notes #

• Ensure your timestamp is sortable: Unix timestamps (numeric) are easiest for comparison, but valid ISO 8601 strings work too (they’re lexicographically comparable).
• Adjust field names (like id or timestamp) in the Map function to match your actual document structure.
• Test with a small dataset first to confirm the View returns the correct latest records before scaling.

内容的提问来源于stack exchange，提问作者willpnw