Uploaded image for project: 'Atlassian Intelligence'
  1. Atlassian Intelligence
  2. AI-1003

Improve Search Functionality for Alphanumeric Terms in Japanese-Embedded Phrases

    • Icon: Suggestion Suggestion
    • Resolution: Unresolved
    • Search - Core
    • None
    • Our product teams collect and evaluate feedback from a number of different sources. To learn more about how we use customer feedback in the planning process, check out our new feature policy.

      Summary

      Currently, Confluence Cloud's search functionality does not return results for alphanumeric terms that are embedded in Japanese text, regardless of the configured indexing language (English, CJK)

      We conducted a test and confirmed that

      • Alphanumeric terms within Japanese text do not appear in search results.
      • The same alphanumeric terms in an English-only phrase are correctly indexed and searchable.
      • This behavior is consistent across all indexing languages (English, CJK). 

      This limitation significantly impacts search accuracy for multilingual teams, as key alphanumeric identifiers (e.g., project codes, IDs) are often embedded in Japanese text and cannot be retrieved efficiently.  

      Reproduction Steps

      1. Create Two Confluence Pages

      We created two separate Confluence pages with the following content:  

      ✅ Page 1 (English-only content)

      WBSe:ABCD1234

      ❌ Page 2 (Japanese + English mixed content)

      チャージ先WBSe:ABCD1234

      2. Search for "ABCD1234"

      We then attempted to search for "ABCD1234" under different indexing language settings. The results were as follows

      Indexing Language Search Query Search Result (Page 1: English-only) Search Result (Page 2: Japanese-mixed)
      English "ABCD1234" ✅ Found ❌ Not Found
      CJK (Japanese) "ABCD1234" ✅ Found ❌ Not Found

      3. Analysis of Search Behavior

      • In all indexing languages, "ABCD1234" was correctly retrieved from the English-only page (Page 1)
      • However, "ABCD1234" inside Japanese text (Page 2) was never found in any indexing language.
      • This indicates that Confluence's search engine does not properly index alphanumeric terms if they are part of a Japanese sentence.

      Key Findings & Issues

      1. Alphanumeric terms in English-only text are indexed and searchable, but the same terms within Japanese text are not.
      2. This issue occurs regardless of the configured indexing language (English, CJK).
      3. Since many organizations use Japanese alongside alphanumeric identifiers (e.g., project codes, ticket numbers), this behavior significantly reduces search usability.

      Expected Behavior & Proposed Improvements

      To improve search functionality, we propose the following enhancements:

      • Alphanumeric terms embedded in Japanese text should be searchable, just like they are in English-only content.
      • Search indexing should recognize alphanumeric sequences inside Japanese phrases, regardless of word boundaries.

            [AI-1003] Improve Search Functionality for Alphanumeric Terms in Japanese-Embedded Phrases

            There are no comments yet on this issue.

              6ed7bc8cd737 Karol Pajak
              9f3b28a42856 Kyungmin Kim
              Votes:
              1 Vote for this issue
              Watchers:
              2 Start watching this issue

                Created:
                Updated: