Redshift cannot parse response from python lambda

Question

I am trying to create and run a lambda function from within Redshift.

I have scaled the lambda_handler back to something very simple

def lambda_handler(event, context):
    print("Received event:", event)

    value = 42
    print("Returning value:", value, "type:", type(value))

    result = [[value]]
    print("JSON result to return:", result)

    return result

I have also tried setting result as 42 and [42].

My create external function statement is

CREATE OR REPLACE EXTERNAL FUNCTION test_fn()
RETURNS INT
VOLATILE
LAMBDA 'arn:aws:.........'
IAM_ROLE 'arn:aws:iam::........';

However I keep getting the error:

SQL Error [XX000]: ERROR: Invalid External Function Response
  Detail: 
  -----------------------------------------------
  error:  Invalid External Function Response
  code:      8001
  context:   Cannot parse External Function response
  query:     0
  location:  exfunc_data.cpp:374
  process:   padbmaster [pid=1073939252]
  -----------------------------------------------

I am using a cluster of 2 ra3.large nodes which is patched to version 191.

I am expecting Redshift to return the value 42 when calling SELECT test_fn();

Pradipta Dash · Accepted Answer · 2025-08-14 10:57:18Z

2

Why Redshift can’t parse your Lambda response — and the minimal fix

Redshift Lambda UDFs must return a JSON string whose top level has a field named results. results must be an array with one element per input row Redshift sends in event["arguments"]. If you return a bare value, a Python list, or anything that isn’t a JSON string with {"results": [...]}, Redshift throws “Cannot parse External Function response”.

Minimal working Python handler (for a no‑arg UDF)

import json

def lambda_handler(event, context):
    # Redshift batches rows in event["arguments"]
    rows = event["arguments"]          # e.g., [ [] ] for SELECT my_fn();
    results = [42] * len(rows)         # one output per input row
    return json.dumps({"results": results})

Checklist

The handler returns json.dumps({...}), not a raw Python object.
Top‑level key is exactly results.
len(return_payload["results"]) == len(event["arguments"]).
Your SQL UDF’s return type matches the Python values you put in results.

If you add parameters later, each “row” in event["arguments"] will be an array of those values; compute one scalar per row and place them in results.

answered Aug 14 at 10:57

Pradipta Dash

362 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

Stephen Saidani Aug 14 at 11:31

Thank you I was missing this, and has resolved my issue.

furas Aug 14 at 13:59

if this answer reolved your problem then you could mark it as accepted, and few minutes later you can upvote it.

Nitiz · Accepted Answer · 2025-08-14 11:21:19Z

Error [XX000]: This error is always about a mismatch between what Lambda returns and what Redshift expects. Once you fix the return type, the error goes away.

Redshift external functions have strict requirements on return types:

The Lambda return type must match the Redshift declared SQL type.
Scalars (INT, VARCHAR, etc.) must be returned as plain values, not as lists, arrays, or dicts.

So, You need to do the following:

value is a plain integer, matching the RETURNS INT type in Redshift.
No brackets, no lists, no dicts — just the scalar.

Lambda Handler:

def lambda_handler(event, context):
    print("Received event:", event)
    
    value = 42
    print("Returning value:", value)
    
    return value  # must be scalar to match RETURNS INT

Redshift External Function

CREATE OR REPLACE EXTERNAL FUNCTION test_fn()
RETURNS INT
VOLATILE
LAMBDA 'arn:aws:lambda:...'
IAM_ROLE 'arn:aws:iam::...';

Calling:

SELECT test_fn();

✅ Returns:

 test_fn
---------
 42

Key Takeaways:

Always match Lambda return to Redshift SQL type.
Scalars: return plain values (42, "hello").
Arrays: return JSON arrays if Redshift type is ARRAY<...>.
Structs/Records: return JSON objects matching the declared schema.
print() statements are fine for debugging — they don’t affect the returned value.

Collectives™ on Stack Overflow

Redshift cannot parse response from python lambda

2 Answers 2

2 Comments

Lambda Handler:

Redshift External Function

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

2 Comments

Lambda Handler:

Redshift External Function

Comments

Your Answer

Sign up or log in

Post as a guest

Related