The variable aggregator cannot correctly aggregate the retrieval results obtained by the knowledge retrieval node

[yangyulun111]

Self Checks

• I have read the Contributing Guide and Language Policy.
• This is only for bug report, if you would like to ask a question, please head to Discussions.
• I have searched for existing issues search for existing issues, including closed ones.
• I confirm that I am using English to submit this report, otherwise it will be closed.
• 【中文用户 & Non English User】请使用英语提交，否则会被关闭 ：）
• Please do not modify this template :) and fill in all the required fields.

Dify version

1.9.1

Cloud or Self Hosted

Self Hosted (Docker)

Steps to reproduce

I have deployed multiple Knowledge Retrieval nodes in Chatflow. I attempted to aggregate them together through a variable aggregator, but the data received and the results output by the variable aggregator are always the retrieval results of the first Knowledge Retrieval node. It cannot receive the results from other retrieval nodes

✔️ Expected Behavior

The variable aggregator integrates all search results and then outputs them

❌ Actual Behavior

The variable aggregator only receives the search results from the first Knowledge Retrieval node and cannot receive the results from other Knowledge Retrieval nodes

[dosubot]

Hi @yangyulun111! I'm Dosu and I’m helping the dify team.

This is a known limitation: the Variable Aggregator node only takes the first available input from upstream branches by design—it does not merge results from multiple Knowledge Retrieval nodes into one output. The proposed aggregate_all feature (which would allow true aggregation) was not merged and is not available in Dify v1.9.1, so you won't see an "Aggregate All" switch in your node settings source, discussion.

For your use case, the recommended workaround is to use a Code node to manually merge the outputs from multiple Knowledge Retrieval nodes. For example, if each node outputs an array, you can use a Code node with something like:

def main(array1, array2):
    merged = (array1 or []) + (array2 or [])
    return {"result": merged}

or in JavaScript:

function main(array1, array2) {
    return { result: (array1 || []).concat(array2 || []) };
}

Then, pass the merged result to your downstream LLM node example.

If you need to format or display the results, you can use a Template node, but this will not preserve structured data for further processing.

The Dify team is working on a feature to allow binding multiple Knowledge Base inputs as context to LLM nodes, which should eventually remove the need for manual merging source.

For more details, see the Variable Aggregator documentation.

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}    

[crazywoola]

https://docs.dify.ai/en/guides/workflow/node/variable-aggregator#variable-aggregator

See the docs, the variable aggregator only aggregates variables not the results. Please read the docs and search the existing issues.