I have found the following Parser() function options effective at removing keys whose values are null. You should set _raw to `JSON.stringify(_raw)` afterwards though.

Yup, I have something similar. My problem is using a tcpjson as your source automatically parses the JSON, and it doesn't create an _raw field. So, I'm not sure how to go about accessing the keys in the event

So when your events hit the pipeline, the JSON object has already been parsed and the contents are extracted as top-level fields (including keys w/null values).

Gotchya, is there a way to access the parsed JSON without re-writing it to _raw & back to the event?

Sorry, my comment was not meant to be a statement. It was meant to be a question.I have a TCP JSON data source on my Stream v3.5.4 instance and events come in as JSON in _raw.So, in my case the TCP JSON source is not automatically parsing the JSON.Which version of Stream are you running?

Looks like I'm on 4.1

and the answer to your question above is yes. There's no _raw key, everything is a top-level field

If there is a specific key you are looking at, you can use that key in parser instead of _raw. However, I assume that you are trying to apply the filter to all fields. I am not sure how to do this without a code function

Do you serialize some of the top-level fields back into _raw as a JSON object at any point in your pipeline?That would be an opportunity to filter out the keys w/null values.