This appears to be a bug and has already been reported on GitHub.
I plan to fix it as is, but it will likely be tomorrow or later, and perhaps the people inside (the project/company) might be faster.
opened 12:32PM - 08 Mar 26 UTC
🐞 bug
### Self Checks
- [x] I have read the [Contributing Guide](https://github.com/l… anggenius/dify/blob/main/CONTRIBUTING.md) and [Language Policy](https://github.com/langgenius/dify/issues/1542).
- [x] This is only for bug report, if you would like to ask a question, please head to [Discussions](https://github.com/langgenius/dify/discussions/categories/general).
- [x] I have searched for existing issues [search for existing issues](https://github.com/langgenius/dify/issues), including closed ones.
- [x] I confirm that I am using English to submit this report, otherwise it will be closed.
- [x] 【中文用户 & Non English User】请使用英语提交,否则会被关闭 :)
- [x] Please do not modify this template :) and fill in all the required fields.
### Dify version
1.13.0, `main`
### Cloud or Self Hosted
Self Hosted (Docker)
### Steps to reproduce
1. Create knowledge and add a document with any metadata (e.g., `hoge = fuga`).
2. Add a Knowledge Retrieval node to the workflow and enable Metadata Filtering.
3. Temporarily set the mode to Automatic and use any model (this is a workaround for a different issue #32372 and PR #32411).
4. Change the mode to Manual and set the Condition to Constant with `hoge` `is` `fuga`, confirming that the knowledge is retrieved as expected.
5, Change from Constant to Variable, and use a variable that contains `fuga` to specify `hoge` `is` `fuga`.
### ✔️ Expected Behavior
The result in step 5 should be the same as in step 4.
### ❌ Actual Behavior
No chunks are retrieved.
Even if we can retrieve chunks in step 4, we can't retrieve any chunks in step 5.