Pipeline reference

# Drop step

A Drop step removes each record where its condition is true, so later steps and destinations never see it.

The condition is an [expression](/docs/code-reference) read on each record.

## Example

The parser declares `level` as `str`.

Step

  Saved JSON

```
{
  "id": "drop",
  "type": "drop",
  "condition": "level == \"debug\""
}
```

Input

```
{
  "level": "info",
  "message": "order placed"
}
```

```
{
  "level": "debug",
  "message": "cache hit"
}
```

```
{
  "level": "error",
  "message": "payment failed"
}
```

→

Output

```
{
  "level": "info",
  "message": "order placed"
}
```

```
{
  "level": "error",
  "message": "payment failed"
}
```

The step drops the debug record. The other two continue in their original order.

## Settings

It also takes the settings [every step](/docs/steps#settings) takes.

`condition`  expression required

Drops the record where it is true. A null condition keeps the record. The GUI's blank is `false`.

Examples:

- `level == "debug"`
- `path == "/health" || user_agent.startsWith("kube-probe")`
- `user.isNull()`

The editor labels `condition` **drop when this is true**.

## How it works

The step reads its condition on each record. A record where it is true leaves the pipeline at this step. Every other record continues unchanged.

Inside an [If step](/docs/steps/if), a Drop step removes only records on that branch.

### Null and missing fields

A condition that is null is not true, so the step keeps the record. A record without `level` passes the same step:

Step

  Saved JSON

```
{
  "id": "drop",
  "type": "drop",
  "condition": "level == \"debug\""
}
```

Input

```
{
  "message": "no level"
}
```

→

Output

```
{
  "message": "no level"
}
```

## See also

-   [Pipeline steps](/docs/steps)
-   [If step](/docs/steps/if)
-   [Expression reference](/docs/code-reference)