Help :How to implement nested loops

[AnZheng333]

Self Checks

• I have searched for existing issues search for existing issues, including closed ones.
• I confirm that I am using English to submit this report (我已阅读并同意 Language Policy).
• [FOR CHINESE USERS] 请务必使用英文提交 Issue，否则会被关闭。谢谢！:)
• Please do not modify this template :) and fill in all the required fields.

1. Is this request related to a challenge you're experiencing? Tell me about your story.

How to implement nested loops

2. Additional context or comments

No response

[dosubot[bot]]

Hi @AnZheng333! I'm Dosu and I’m helping the dify team.

You can implement nested loops in Dify workflows by placing a Loop node or an Iteration node inside the sub-workflow of another Loop or Iteration node. This allows you to create complex, multi-level repetition structures—useful for tasks like processing multi-dimensional data or performing repeated refinements.

• The Loop node executes a sub-workflow multiple times, supporting stateful, progressive refinement. You can control it with termination conditions, a maximum loop count, or an Exit Loop node. Each iteration can use and update loop variables, and you can nest any node—including another Loop or Iteration node—inside its body.
• The Iteration node processes arrays by running a sub-workflow for each element, either sequentially or in parallel. You can nest Iteration nodes for multi-dimensional array processing, or place a Loop node inside an Iteration node for repeated operations per item.
• Both nodes treat their sub-workflows as subgraphs, so nesting is fully supported in both the backend and UI. Loop variables and break conditions are tracked and updated across iterations, even in nested scenarios.

Example:
Suppose you want to process a list of chapters, and for each chapter, refine its content multiple times. You would use an Iteration node to loop over chapters, and inside its sub-workflow, add a Loop node to refine each chapter.

Here’s a Mermaid diagram to illustrate:

graph TD
    A["Start"] --> B["Iteration: For each chapter"]
    B --> C["Loop: Refine chapter content"]
    C --> D["LLM: Generate/Refine"]
    D --> E["Exit Loop (if done)"]
    E --> F["End Iteration"]
    F --> G["End"]

Loading

For more details and examples, see the Dify docs on Loop nodes and Iteration nodes.

If you’re looking for a general programming example, a nested loop in Python looks like this:

for i in range(3):
    for j in range(2):
        print(f"Outer: {i}, Inner: {j}")

Let me know if you need a step-by-step guide for a specific use case!

To reply, just mention @dosu.

---

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

   

[AnZheng333]

Hello, thank you for your reply. However, my previous approach was to directly add a child loop node inside the loop body, but I found that it wasn't possible to add one, just like in this image.Or do I need to use Iteration nodes and Loop nodes together to achieve the nested effect?

[dosubot[bot]]

You can't directly add a Loop node inside another Loop node in the Dify workflow editor UI. The UI intentionally prevents this—when you're inside a Loop (or Iteration) node, the option to add another Loop or Iteration node is filtered out and won't appear in the node selection menu. This is enforced by the editor's code and matches what you saw in your screenshot and experience (source).

Currently, there isn't a built-in way to nest Loop or Iteration nodes directly inside each other in the visual editor. If you need nested repetition, you'll need to design your workflow by chaining separate loop/iteration scopes, not by embedding one inside another (source). If you have a specific use case, I can help suggest a workaround or alternative design!

To reply, just mention @dosu.

---

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

   

[hjlarry]

try workflow_as_tool

[heathdutton]

For nested loops, use the Iteration node with a nested structure:

1. Add an Iteration node for your outer loop - feed it an array
2. Inside that iteration, add another Iteration node for the inner loop
3. Each iteration has access to the current item via the iterator variable

Alternatively, you can flatten nested data in a Code node first:

def main(outer_list: list) -> dict:
    flattened = []
    for outer in outer_list:
        for inner in outer["items"]:
            flattened.append({"outer": outer, "inner": inner})
    return {"result": flattened}

Then iterate over the flattened list once.

Which approach fits your use case better?