$lookup
Definition
Adds matching records from a host-provided in-memory dataset.
Core aggregation stage · Combine pipelines. Use this syntax in a Core pipeline; Studio availability depends on the Flow compiler.
Syntax
Basic form, with concrete values:
{
"$lookup": {
"from": [
{
"sku": "P1",
"price": 5
}
],
"localField": "sku",
"foreignField": "sku",
"as": "products"
}
}
Parameters
| Parameter | Required | Description |
|---|---|---|
from |
Yes | Array/object of foreign records, or a dataset name provided in execution globals under foreignCollections. |
localField |
Yes | Local join field path. |
foreignField |
Yes | Foreign join field path. |
as |
No | Output field containing matched foreign records. Defaults to results. |
cacheOptions |
No | Cache settings; maxEntries defaults to 1024 distinct keys. Size the host dataset accordingly. |
Behavior
- There is no database connection or collection read. The host supplies all records.
- No match yields an empty array. Lookup uses key equality in the in-memory index.
- An index stops loading after its configured key limit; large datasets can therefore omit matches.
- A named dataset must be present even when input is empty.
Empty input and invalid specification
These cases use the basic form above unless the specification is explicitly replaced.
| Case | Observed result |
|---|---|
| empty input | [] |
invalid specification (null) |
Error: IllegalArgumentException — $lookup stage expects an object specification |
Examples
Basic pipeline
Input documents:
[
{
"sku": "P1"
},
{
"sku": "P2"
}
]
Pipeline:
[
{
"$lookup": {
"from": [
{
"sku": "P1",
"price": 5
}
],
"localField": "sku",
"foreignField": "sku",
"as": "products"
}
}
]
Result:
[
{
"sku": "P1",
"products": [
{
"sku": "P1",
"price": 5
}
]
},
{
"sku": "P2",
"products": []
}
]
Missing named dataset
Input documents:
[
{
"sku": "P1"
}
]
Pipeline:
[
{
"$lookup": {
"from": "products",
"localField": "sku",
"foreignField": "sku",
"as": "products"
}
}
]
Observed error: foreign collection 'products' not available
Verification
The 4 examples on this page were executed through Core PipelineExecutor with a fresh context for each case. They establish the displayed behavior, not all possible option combinations.
Download the executable cases and recorded results.
MongoDB differences
Stage examples were executed in Core. They were not compared live with MongoDB in the expression parity suite.
- Database collection access, correlated lookup pipelines and MongoDB index behavior are not implemented.
MongoDB $lookup reference. Read the Core behavior above before reusing MongoDB examples.