I have a simple json
{
"years": [ "2018", "2019" ]
}
that I want to read with Benthos. However as this is a "beautified" json, Benthos reads each line as message. How can one read this multiline json with benthos and compactify it for further downstream processors?
input:
file:
paths: [./init.json]
codec: lines/multipart
pipeline:
processors: ??
output:
type: stdout
The pipeline should simply print the same in stdout
as jq -c ./init.json
would.
try this:
input:
file:
paths: [./init.json]
codec: all-bytes
processors:
- mapping: |
root = this.format_json()
output:
stdout:
codec: all-bytes
Also, note that the current release is V4, where type: stdout
isn't how you specify the output. Details here.