Multiple Row Functions

NOTE

Verification Status: Tested and Verified against rekuiper engine with live telemetry stream load on 2026-10-01 17:56:45 UTC.
Scorecard: 2 Verified, 0 Unsupported, 0 Broken:

• Verified (2): unnest, extract.

A multiple-row function expands input data structures into multiple output rows.

You can use multiple-row functions only in the SELECT clause of a query. The engine currently permits only one multiple-row function per SELECT clause.

UNNEST

unnest(array)

Expands an array into multiple output rows. The argument must evaluate to an array.

When array elements are objects (map[string]interface{}), the function converts the object fields into distinct columns in the resulting rows.

Examples

Create a stream named demo with the following input record:

{
  "a": [
    1,
    2
  ],
  "b": 3
}

Rule to unnest array values:

SQL: SELECT unnest(a) FROM demo
___________________________________________________
{"unnest":1}
{"unnest":2}

Rule to unnest array values with adjacent columns:

SQL: SELECT unnest(a), b FROM demo
___________________________________________________
{"unnest":1, "b":3}
{"unnest":2, "b":3}

Create a stream named demo with objects inside an array:

{
  "x": [
    {
      "a": 1,
      "b": 2
    },
    {
      "a": 3,
      "b": 4
    }
  ],
  "c": 5
}

Rule to unnest array objects into columns:

SQL: SELECT unnest(x), c FROM demo
___________________________________________________
{"a":1, "b":2, "c": 5}
{"a":3, "b":4, "c": 5}

EXTRACT

extract(map[string]interface{})

Expands a map or JSON object and flattens its key-value pairs into individual columns on the current row. The argument must evaluate to a map object.

Example

Create a stream named demo with the following input record:

{
  "data": {
    "k1": "v1",
    "k2": "v2"
  }
}

Rule to extract map fields into row columns:

SQL: SELECT extract(data) FROM demo
__________________________________________________
{"k1":"v1","k2":"v2"}