drop_null_fields

Returns a record with fields whose value is null removed.

drop_null_fields(x:record, field:field...) -> record

Description

The drop_null_fields function returns a copy of x with every field whose value is null removed. Without explicit selectors, every field is considered, including fields nested inside record-typed values. With selectors, only the named fields (and any fields nested within them) are considered for removal; other null fields are preserved.

Unlike print_ndjson(strip_null_fields=true), the function never serializes the record, so secret-typed values flow through unchanged and reach the sink unredacted.

`x: record`

The record from which to remove null fields.

`field: field... (optional)`

A comma-separated list of field paths inside x to consider. When omitted, every field is considered.

Examples

Drop all null fields

from {
  id: 42,
  status: "active",
  comment: null,
}
this = drop_null_fields(this)

{
  id: 42,
  status: "active",
}

Build an HTTP request body inline

A common use case is cleaning optional null fields out of an inline body record before sending it to an API that rejects JSON null values:

from_http "https://api.example.com/v1/lookup",
  method="POST",
  body=drop_null_fields({
    version: 1,
    query: "example",
    options: null,
  })

The wire body contains only version and query; options is omitted.

Restrict the drop to specific fields

from {
  a: null,
  b: null,
  c: null,
  d: 1,
}
this = drop_null_fields(this, a, c)

{
  b: null,
  d: 1,
}

b stays in the result because it was not named in the selector list, even though it is null.

Recursive drop inside nested records

from {
  metadata: {
    created: "2024-01-01",
    updated: null,
  },
  data: {
    value: 42,
    comment: null,
  },
}
this = drop_null_fields(this)

{
  metadata: {
    created: "2024-01-01",
  },
  data: {
    value: 42,
  },
}