langchain-ai/langchain · error · ValueError

Saving an example selector is not currently supported

Error message

Saving an example selector is not currently supported

What it means

`FewShotPromptWithTemplates.save()` raises this `ValueError` when the prompt has an `example_selector` set. Serialization to disk only supports prompts whose examples are inlined; example selectors (e.g. `LengthBasedExampleSelector`, `SemanticSimilarityExampleSelector`) hold arbitrary state (vector stores, embedders) that the save format cannot represent.

Source

Thrown at libs/core/langchain_core/prompts/few_shot_with_templates.py:238

    @deprecated(
        since="1.2.21",
        removal="2.0.0",
        alternative="Use `dumpd`/`dumps` from `langchain_core.load` to serialize "
        "prompts and `load`/`loads` to deserialize them.",
    )
    def save(self, file_path: Path | str) -> None:
        """Save the prompt to a file.

        Args:
            file_path: The path to save the prompt to.

        Raises:
            ValueError: If `example_selector` is provided.
        """
        if self.example_selector:
            msg = "Saving an example selector is not currently supported"
            raise ValueError(msg)
        return super().save(file_path)

View on GitHub (pinned to e32fa9a52e)

Solutions

1. Drop the selector before saving: rebuild a `FewShotPromptWithTemplates` with an explicit `examples=[...]` list and save that.
2. Serialize with `langchain_core.load.dumpd`/`dumps` if the selector itself is serializable, and reload via `load`/`loads`.
3. Persist the selector's configuration separately (e.g. the examples list plus selector type/params) and reconstruct the prompt at load time.

Example fix

// before
prompt = FewShotPromptTemplateWithTemplates(
    example_selector=LengthBasedExampleSelector(...),
    ...
)
prompt.save('few_shot.json')  # ValueError

// after
prompt = FewShotPromptWithTemplates(
    examples=examples,  # inline list instead of selector
    ...
)
prompt.save('few_shot.json')

# before saving, check selector presence
if getattr(prompt, 'example_selector', None):
    raise ValueError(
        'Cannot save prompt with an example_selector; inline examples first.'
    )
prompt.save('prompt.json')

from langchain_core.prompts.few_shot_with_templates import FewShotPromptWithTemplates

def is_savable_few_shot(p) -> bool:
    return isinstance(p, FewShotPromptWithTemplates) and p.example_selector is None

try:
    prompt.save(path)
except ValueError as e:
    if 'example selector' in str(e).lower():
        # fall back: save a selector-free copy
        FewShotPromptWithTemplates(
            examples=prompt.examples, example_prompt=prompt.example_prompt,
            suffix=prompt.suffix, prefix=prompt.prefix,
            input_variables=prompt.input_variables,
        ).save(path)
    else:
        raise

When it happens

Trigger: Calling `prompt.save('file.json')` (or `prompt.save_prompt`) on a `FewShotPromptWithTemplates` instance constructed with a non-None `example_selector` argument.

Common situations: Building a few-shot prompt with dynamic example selection, then trying to persist it to YAML/JSON for deployment; migrating older LangChain code that saved plain `FewShotPromptTemplate` objects and assuming selectors also serialize.

Related errors

• Saving an example selector is not currently supported
• Cannot save prompt with partial variables.
• Prompt {self} does not support saving.
• {save_path} must be json or yaml
• Only one of 'examples' and 'example_selector' should be prov

AI-assisted analysis of langchain-ai/langchain@e32fa9a52e (2026-08-14). Data as JSON: /api/errors/babceee8bf242cbb. Report an issue: GitHub.