redis/redis-py · error · DataError

Invalid input of type

Error message

Invalid input of type: '{typename}'. Convert to a bytes, string, int or float first.

What it means

Raised by Encoder.encode() (redis/_parsers/encoders.py:29) when the value is not one of bytes/bytearray/memoryview/bool/int/float/str. Common culprits are None, list, dict, tuple, and custom objects. The encoder only handles the primitive RESP-serializable types, so anything else raises DataError naming the offending type.

Solutions

  1. Serialize complex types before sending: json.dumps(obj).encode() or another codec.
  2. Convert None to an explicit sentinel bytes value if None is meaningful.
  3. For lists, spread elements across the right command (RPUSH key *items) rather than passing one list arg.
  4. Add a pre-send normalization layer so only primitives reach the client.

Example fix

# before
r.set("user", {"id": 1, "name": "x"})  # DataError: Invalid input of type: 'dict'

# after
import json
r.set("user", json.dumps({"id": 1, "name": "x"}))
Defensive patterns

Strategy: type-guard

Validate before calling

# Reject non-primitive values before they reach the encoder
def encode_safe(v):
    if not isinstance(v, (bytes, bytearray, memoryview, int, float, str)):
        raise TypeError(f"serialize {type(v).__name__} before sending to Redis")
    return v
r.set("k", encode_safe(json.dumps(payload)))

Type guard

def is_encodable(v) -> bool:
    return isinstance(v, (bytes, bytearray, memoryview, str, int, float)) and not isinstance(v, bool)

Try / catch

from redis.exceptions import DataError
try:
    r.set("k", value)
except DataError as e:
    if "Invalid input of type" in str(e):
        r.set("k", json.dumps(value))

Prevention

When it happens

Trigger: r.set('k', None), r.set('k', [1,2,3]), r.set('k', {'a':1}), or passing a dataclass/object instance directly. Any arg routed through the encoder that isn't a primitive triggers it.

Common situations: Forgetting to serialize complex types; passing None where a value is expected; treating Redis like a Python object store without a codec; list-as-single-arg instead of spreading elements (e.g. RPUSH needs *items).

Related errors


AI-assisted analysis of redis/redis-py@6a6b581b48 (2026-08-10). Data as JSON: /api/errors/4bcb2aed0204a1bd. Report an issue: GitHub.

Appendix: source

Thrown at redis/_parsers/encoders.py:29

        self.encoding_errors = encoding_errors
        self.decode_responses = decode_responses

    def encode(self, value):
        "Return a bytestring or bytes-like representation of the value"
        if isinstance(value, (bytes, bytearray, memoryview)):
            return value
        elif isinstance(value, bool):
            # special case bool since it is a subclass of int
            raise DataError(
                "Invalid input of type: 'bool'. Convert to a "
                "bytes, string, int or float first."
            )
        elif isinstance(value, (int, float)):
            value = repr(value).encode()
        elif not isinstance(value, str):
            # a value we don't know how to deal with. throw an error
            typename = type(value).__name__
            raise DataError(
                f"Invalid input of type: '{typename}'. "
                f"Convert to a bytes, string, int or float first."
            )
        if isinstance(value, str):
            value = value.encode(self.encoding, self.encoding_errors)
        return value

    def decode(self, value, force=False):
        "Return a unicode string from the bytes-like representation"
        if self.decode_responses or force:
            if isinstance(value, memoryview):
                value = value.tobytes()
            if isinstance(value, bytes):
                value = value.decode(self.encoding, self.encoding_errors)
        return value

View on GitHub (pinned to 6a6b581b48)