Run a Map-Reduce job

Map-Reduce runs a server-side JavaScript job on a collection: map() emits key and value pairs, reduce() combines the values of each key, and an optional finalize() shapes the result. The editor lets you preview every step on a sample before running the job on the whole collection.

MongoDB deprecates map-reduce since version 5.0, though it still runs it. For new work, the Aggregation Editor is usually the better choice.

Run a job

1. Select a collection in the connection tree.
2. Press Ctrl+M, or choose Collection > Open Map-Reduce. A new tab opens with default functions; output is inline.
3. Write the map() function:, for example function () { emit(this.status, this.total); }.
4. Write the reduce(key, values) function:, for example function (key, values) { return Array.sum(values); }.
5. Optionally tick Use filter/sort/limit and fill Query:, Sort: and Limit:, tick Use finalize for a finalize(key, reducedValue) function:, or Use global scope for Global variables, as a document:.
6. Preview the steps (below), then click Execute (Run the Map-Reduce job on the whole collection). The result opens in the Finished job tab.

Preview each step

Previews never write: they run on the first 100 documents of the input (fewer if Limit says so), always with inline output.

Output options

Writing into a collection asks first (Execute Map-Reduce: The job writes its result to … Run it?) and is refused on a read-only connection. Replace needs a collection name.

Options also offers jsMode and verbose, passed to the server as they are.

Files and details

Result tabs are read-only, with the Table, Tree and JSON views and their copy options.

Troubleshooting