bug(core): _batch and _abatch lack batch_size validation causing infinite loop

[chang923]

Checked other resources

• This is a bug, not a usage question.
• I added a clear and descriptive title that summarizes this issue.
• I used the GitHub search to find a similar question and didn't find it.
• I am sure that this is a bug in LangChain rather than my code.
• The bug is not resolved by updating to the latest stable version of LangChain (or the specific integration package).
• This is not related to the langchain-community package.
• I posted a self-contained, minimal, reproducible example. A maintainer can copy it and run it AS IS.

Package (Required)

• langchain
• langchain-openai
• langchain-anthropic
• langchain-classic
• langchain-core
• langchain-model-profiles
• langchain-tests
• langchain-text-splitters
• langchain-chroma
• langchain-deepseek
• langchain-exa
• langchain-fireworks
• langchain-groq
• langchain-huggingface
• langchain-mistralai
• langchain-nomic
• langchain-ollama
• langchain-openrouter
• langchain-perplexity
• langchain-qdrant
• langchain-xai
• Other / not sure / general

Related Issues / PRs

#36648

Reproduction Steps / Example Code (Python)

from langchain_core.indexing.api import _batch, _abatch

# Synchronous _batch with size=0 leads to infinite loop
list(_batch(0, [1, 2, 3]))

# Asynchronous _abatch with size=0 yields empty batches endlessly
import asyncio
from typing import AsyncIterable

async def async_iter(data):
    for item in data:
        yield item

async def test_abatch():
    async for batch in _abatch(0, async_iter([1, 2, 3])):
        print(batch)  # Prints empty lists indefinitely

asyncio.run(test_abatch())

Error Message and Stack Trace (if applicable)

Description

I am using aindex and index functions from langchain_core.indexing. When the batch_size parameter is mistakenly set to 0 (or any negative value), the internal utility functions _batch and _abatch exhibit harmful behavior:

• Synchronous _batch: Enters an infinite loop (while True never exits because islice(it, 0) returns empty list indefinitely).
• Asynchronous _abatch: Yields an empty list for every element in the input iterable, causing silent performance degradation and wasted CPU.

Expected behavior:
The functions should raise a ValueError immediately when size <= 0, failing fast and alerting the user to the invalid configuration.

Actual behavior:
_batch hangs the process; _abatch produces a flood of empty batches without any error.

Proposed fix:
Add a simple validation at the beginning of both functions:

if size <= 0:
    raise ValueError("Batch size must be a positive integer.")

### System Info

System Information
------------------
> OS:  Linux
> OS Version:  #1 SMP Tue Nov 5 00:21:55 UTC 2024
> Python Version:  3.10.20 (main, Mar 11 2026, 17:46:40) [GCC 14.3.0]

Package Information
-------------------
> langchain_core: 1.2.19
> langchain: 1.2.13
> langchain_community: 0.4.1
> langsmith: 0.7.17
> langchain_chroma: 1.1.0
> langchain_classic: 1.0.3
> langchain_deepseek: 1.0.1
> langchain_huggingface: 1.2.1
> langchain_openai: 1.1.11
> langchain_text_splitters: 1.1.1
> langgraph_sdk: 0.3.11

Optional packages not installed
-------------------------------
> deepagents
> deepagents-cli

Other Dependencies
------------------
> aiohttp: 3.13.3
> async-timeout: 4.0.3
> chromadb: 1.5.5
> dataclasses-json: 0.6.7
> httpx: 0.28.1
> httpx-sse: 0.4.3
> huggingface-hub: 1.7.1
> jsonpatch: 1.33
> langgraph: 1.1.2
> numpy: 2.2.6
> openai: 2.28.0
> opentelemetry-api: 1.40.0
> opentelemetry-sdk: 1.40.0
> orjson: 3.11.7
> packaging: 25.0
> pydantic: 2.12.5
> pydantic-settings: 2.13.1
> pytest: 9.0.3
> pyyaml: 6.0.3
> PyYAML: 6.0.3
> requests: 2.32.5
> requests-toolbelt: 1.0.0
> rich: 14.3.3
> sentence-transformers: 5.3.0
> SQLAlchemy: 2.0.48
> sqlalchemy: 2.0.48
> tenacity: 9.1.4
> tiktoken: 0.12.0
> tokenizers: 0.22.2
> transformers: 5.3.0
> typing-extensions: 4.15.0
> uuid-utils: 0.14.1
> websockets: 16.0
> xxhash: 3.6.0
> zstandard: 0.25.0

[chang923]

Eugene Yurtsev (@eyurtsev) Could you please assign this issue to me? I have a PR ready. Thanks!

[jackjin1997]

Hit the same thing yesterday while debugging a slow indexing pipeline — turned out the batch_size config was accidentally zeroed and _batch just hung forever. The fix is straightforward: add a ValueError guard at the top of both _batch and _abatch when size <= 0. Could I be assigned?

[sharziki]

Hi! I'd like to take this one. The fix is straightforward — add a batch_size <= 0 validation at the top of both _batch and _abatch in langchain_core/indexing/api.py to raise a ValueError instead of entering an infinite loop. Happy to submit a PR with tests once assigned.

[meridianmindx]

Excellent bug report with clear root cause analysis. The issue is that _batch and _abatch don't validate batch_size before entering the processing loop, leading to infinite loops or silent performance degradation.

The proposed fix (validate size <= 0 and raise ValueError) is straightforward and follows Pythonic "fail fast" principles. This is a good catch because such bugs can be very subtle in production.

One additional suggestion: consider adding a test case that verifies the error message is helpful to users (e.g., "Batch size must be a positive integer, got 0").

Thanks for finding this edge case and providing a minimal reproducer!

[hijingsong]

Looking into this — adding a batch_size <= 0 guard in _batch and _abatch to raise ValueError early. Happy to submit a fix.