pola-rs/polars · error · ValueError

{credential_provider}

Error message

{credential_provider}

What it means

`pl.Config.set_default_credential_provider(provider)` only accepts the string 'auto' (reset to automatic detection) or an actual credential provider object such as `pl.CredentialProviderAWS(...)`. Any other string raises ValueError whose message is just the offending string itself — an unusual pattern, so the traceback looks odd but simply means 'invalid credential provider name'.

Source

Thrown at py-polars/src/polars/config.py:1732

        credential_provider
            Provide a function that can be called to provide cloud storage
            credentials. The function is expected to return a dictionary of
            credential keys along with an optional credential expiry time.

            Can also be set to None, which globally disables auto-initialization
            of credential providers, or "auto" (the default behavior).

        Examples
        --------
        >>> pl.Config.set_default_credential_provider(
        ...     pl.CredentialProviderAWS(
        ...         assume_role={"RoleArn": "...", "RoleSessionName": "..."}
        ...     )
        ... )
        <class 'polars.config.Config'>
        """
        if isinstance(credential_provider, str) and credential_provider != "auto":
            raise ValueError(credential_provider)

        _set_default_credential_provider(credential_provider)

        return cls

    @classmethod
    def reload_env_vars(cls) -> None:
        """
        Update the Polars config from the set environment variables.

        Normally you need not call this, it is only necessary when updating
        undocumented (environment-variable-only) config flags from Python after
        importing Polars.
        """
        plr.config_reload_env_vars()

View on GitHub (pinned to 68506541d2)

Solutions

1. Pass an instance: `pl.Config.set_default_credential_provider(pl.CredentialProviderAWS(assume_role={...}))`.
2. Pass 'auto' to restore automatic credential resolution, or None to clear the override.
3. Keep provider selection in Python code rather than serializing provider names to config files.

Example fix

# before
pl.Config.set_default_credential_provider("aws")  # ValueError: aws

# after
pl.Config.set_default_credential_provider(
    pl.CredentialProviderAWS(assume_role={"RoleArn": "...", "RoleSessionName": "..."})
)

def set_credential_provider(provider: object) -> None:
    if isinstance(provider, str) and provider != "auto":
        raise ValueError(
            f"credential provider must be an instance or 'auto', got name {provider!r}; "
            "e.g. pl.CredentialProviderAWS(...)"
        )
    pl.Config.set_default_credential_provider(provider)

def is_valid_credential_provider(provider: object) -> bool:
    return not isinstance(provider, str) or provider == "auto"

When it happens

Trigger: `pl.Config.set_default_credential_provider('aws')` or 'auto_detect' — passing a provider name where an object is required; loading provider choice from a config file as a string.

Common situations: Assuming providers are selectable by name string; migrating code that stored provider names; mixing up this global default with the per-call credential_provider argument of scan/read functions.

Related errors

• `separator` must be a single character; found {separator!r}
• number of characters must be > 0
• invalid alignment: {format!r}
• invalid table format name: {format!r} Expected one of: {', '
• limit should be positive

AI-assisted analysis of pola-rs/polars@68506541d2 (2026-08-19). Data as JSON: /api/errors/cfc22a6d9c2881f2. Report an issue: GitHub.