langgenius/dify
1
0
Fork 0
mirror of https://github.com/langgenius/dify.git synced 2026-07-22 12:36:14 -04:00
Code Issues 775 Packages Projects Releases 10 Wiki Activity

How to retrieve a whole document from knowledge base? #11215

New Issue
Closed
opened 2026-02-21 18:56:59 -05:00 by yindo · 0 comments
yindo commented 2026-02-21 18:56:59 -05:00
Owner
Copy Link
Copy Source

Originally created by @kandeng on GitHub (Mar 16, 2025).

Self Checks

  • 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 (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.

Dify version

1.0.0

Cloud or Self Hosted

Self Hosted (Source)

Steps to reproduce

  1. Create a knowledge base, named as "our_products",

Upload a few product introduction DOCX files to "our_products", including a file named "JCLD003010N001-W2.docx".

  1. Do chunking for the files in the knowledge base "our_products",

Since the maximum chunk size is quite small, 4000 tokens, we selected 2048 token/chunk.

  1. Create a "chatbot" instance in the studio, mainly consists of 2 steps,

3.1 With user's query, retrieve the files from the "our_products" knowledge base,

3.2 Input both the files from knowledge base, and the user's query, to the LLM that is a DS-R1, asking R1 to read and comprehend the file's content.

  1. Our query is "Please tell me the size of the product called 'JCLD003010N001-W2'? ",

    We expected the R1 would give an answer like "35*57",

    however, the R1 replied "I don't know".

  2. We dived into the workflow step by step, found the root cause, that is,

Step 3.1 should give step 3.2 a whole file's content,

However, maybe because of chunking, what step 3.1 really gave step 3.2 was a string, "JCLD003010N001-W2specification", rather than the entire file content.

✔️ Expected Behavior

"Knowledge retrieval" block is expected to retrieve the entire contents of the files in the knowledge base.

❌ Actual Behavior

"Knowledge retrieval" block only retrieves one chunk from the file content in the knowledge base.

Originally created by @kandeng on GitHub (Mar 16, 2025). ### Self Checks - [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 (我已阅读并同意 [Language Policy](https://github.com/langgenius/dify/issues/1542)). - [x] [FOR CHINESE USERS] 请务必使用英文提交 Issue，否则会被关闭。谢谢！:） - [x] Please do not modify this template :) and fill in all the required fields. ### Dify version 1.0.0 ### Cloud or Self Hosted Self Hosted (Source) ### Steps to reproduce 1. Create a knowledge base, named as "our_products", Upload a few product introduction DOCX files to "our_products", including a file named "JCLD003010N001-W2.docx". 2. Do chunking for the files in the knowledge base "our_products", Since the maximum chunk size is quite small, 4000 tokens, we selected 2048 token/chunk. 3. Create a "chatbot" instance in the studio, mainly consists of 2 steps, 3.1 With user's query, retrieve the files from the "our_products" knowledge base, 3.2 Input both the files from knowledge base, and the user's query, to the LLM that is a DS-R1, asking R1 to read and comprehend the file's content. 4. Our query is "Please tell me the size of the product called 'JCLD003010N001-W2'? ", We expected the R1 would give an answer like "35*57", however, the R1 replied "I don't know". 5. We dived into the workflow step by step, found the root cause, that is, Step 3.1 should give step 3.2 a whole file's content, However, maybe because of chunking, what step 3.1 really gave step 3.2 was a string, "JCLD003010N001-W2specification", rather than the entire file content. ### ✔️ Expected Behavior "Knowledge retrieval" block is expected to retrieve the entire contents of the files in the knowledge base. ### ❌ Actual Behavior "Knowledge retrieval" block only retrieves one chunk from the file content in the knowledge base.
yindo closed this issue 2026-02-21 18:56:59 -05:00
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
Knowledge Pipeline
Proposal
WIP
cloud
dependencies
feat:webapp
good first issue
outdated
performance
pull-request
Mirrored from GitHub Pull Request
 python:uv
refactor
size:L
stale
status: accepting prs
status: blocked
status: in discussion
stt & tts
vulnerablity
web
wont-fix
🌊 feat:workflow
🌚 invalid
🐞 bug
👻 feat:rag
💪 enhancement
📚 documentation
🔨 feat:tools
🙋‍♂️ question
🙏 help wanted
🤔 cant-reproduce
🤖 feat:agent
No Label
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
yindo
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: langgenius/dify#11215
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?