Aggregation editor

Open the aggregation editor with Mod+Shift+A, or convert a find from the query bar.

Stages

• Add, reorder, disable and delete stages. A disabled stage stays in the pipeline but is skipped, which is handy when you're narrowing down a problem.
• Preview stage by stage: select a stage to see the documents coming out of it, so you can check each step before running the whole pipeline.
• Each stage is edited in its own code editor.

Operators

The operator library covers the full aggregation language, including $lookup, $facet, $graphLookup, $unionWith, $setWindowFields, $densify, $fill, and Atlas's $search and $vectorSearch.

Options

• Allow disk use for large sorts and groups.
• maxTimeMS to cap the run time on the server.
• Collation.

Templates

Start from a built-in template or save your own pipelines as templates.

Writes

$out and $merge write to a collection, so they count as writes: they are blocked on read-only connections, including inside explain.

Next steps

• Query Code: generate the pipeline in mongo shell, Node.js, Python, Java, C#, PHP, Ruby, Go or Rust.
• Explain the pipeline to see which stages use indexes. See Explain plans.
• AI debugging: with AI allowed on the connection, the assistant can walk through a pipeline stage by stage. See AI assistant.