langchain-ai/langchain · error · ValueError

Only one of example_prompt and example_prompt_path should be

Error message

Only one of example_prompt and example_prompt_path should be specified.

What it means

In `_load_few_shot_prompt`, the example prompt may be given inline (`example_prompt: {...config...}`) or by file (`example_prompt_path: ...`), but not both — supplying both keys raises this `ValueError`, mirroring the template/template_path exclusivity rule.

Source

Thrown at libs/core/langchain_core/prompts/loading.py:170

def _load_few_shot_prompt(
    config: dict[str, Any], *, allow_dangerous_paths: bool = False
) -> FewShotPromptTemplate:
    """Load the "few shot" prompt from the config."""
    # Load the suffix and prefix templates.
    config = _load_template(
        "suffix", config, allow_dangerous_paths=allow_dangerous_paths
    )
    config = _load_template(
        "prefix", config, allow_dangerous_paths=allow_dangerous_paths
    )
    # Load the example prompt.
    if "example_prompt_path" in config:
        if "example_prompt" in config:
            msg = (
                "Only one of example_prompt and example_prompt_path should "
                "be specified."
            )
            raise ValueError(msg)
        example_prompt_path = Path(config.pop("example_prompt_path"))
        if not allow_dangerous_paths:
            _validate_path(example_prompt_path)
        config["example_prompt"] = load_prompt(
            example_prompt_path, allow_dangerous_paths=allow_dangerous_paths
        )
    else:
        config["example_prompt"] = load_prompt_from_config(
            config["example_prompt"], allow_dangerous_paths=allow_dangerous_paths
        )
    # Load the examples.
    config = _load_examples(config, allow_dangerous_paths=allow_dangerous_paths)
    config = _load_output_parser(config)
    return FewShotPromptTemplate(**config)


def _load_prompt(
    config: dict[str, Any], *, allow_dangerous_paths: bool = False

View on GitHub (pinned to e32fa9a52e)

Solutions

1. Delete one key: keep the inline `example_prompt` config dict, or keep `example_prompt_path` pointing at another prompt JSON/YAML file.
2. Add a config lint step that rejects files containing both keys before loading.

Example fix

# before
# {"_type": "few_shot", "example_prompt": {...},
#  "example_prompt_path": "ep.json", ...}
load_prompt('prompt.json')  # ValueError

# after
# {"_type": "few_shot", "example_prompt": {"_type": "prompt",
#   "template": "Q: {q}\nA: {a}", "input_variables": ["q", "a"]}, ...}
load_prompt('prompt.json')

if 'example_prompt' in config and 'example_prompt_path' in config:
    del config['example_prompt_path']  # inline wins; or raise
load_prompt_from_config(config)

def has_no_example_prompt_conflict(config: dict) -> bool:
    return not ('example_prompt' in config and 'example_prompt_path' in config)

When it happens

Trigger: A few-shot config JSON containing both `"example_prompt"` and `"example_prompt_path"` keys, loaded via `load_prompt`.

Common situations: Editing a working config and pasting in an example-prompt config without deleting the old path key; merging prompt configs from two sources.

Related errors

• Invalid examples format. Only list or string are supported.
• Both `{var_name}_path` and `{var_name}` cannot be provided.
• Loading {config_type} prompt not supported
• Unsupported template file format: '{resolved_path.suffix}'.
• Invalid file format. Only json or yaml formats are supported

AI-assisted analysis of langchain-ai/langchain@e32fa9a52e (2026-08-14). Data as JSON: /api/errors/c8c1c6861060ef90. Report an issue: GitHub.