Why does the n8n Python CODE node omit keys with null values in JSON output?

The issue is that your Python CODE node isn’t outputting the keys with null values even though your Python dictionary has them set as None. In other words, when you do:

PYTHONCopied
return [ final_json_dict ]

the resulting JSON output omits any keys whose value is None (which is the Python equivalent of JSON null) instead of keeping them—yet your JSON schema requires those keys to be present with a value of null.

A couple of points to consider:

1. In n8n each output item is normally expected to be wrapped in a key called json (for example: { "json": { ... } }). While n8n does do some automatic wrapping in certain nodes (like the Function or Code nodes), if the output isn’t structured exactly as expected the conversion may drop keys with None values.

2. The behavior you’re observing isn’t causing an error message; it’s simply that during the conversion/serialization process the keys whose values evaluate to None are not retained in the final output, which means your returned object no longer fully complies with the required JSON schema.

To work around this, make sure your node returns data in the expected structure. For example:

PYTHONCopied
return [
    {
        "json": final_json_dict
    }
]

This way, n8n has a clear wrapper for your data. If the node (or the conversion layer) still omits keys with null values and you need those keys explicitly present, you might need to revisit how the conversion is done or add a step to explicitly inject nulls into the final JSON output. Unfortunately, as of now, n8n’s handling for this case in the Code node (when running on Cloud) will drop keys with null values during serialization.

In summary, the “problem” is that n8n’s output processing for the CODE node (Python in your case) is stripping out keys with None (null) values, so your final output doesn’t match the JSON schema that requires them.

Let me know if you need further assistance with this!