JIS Function (Gnumeric)
Convert selected half-width text with Gnumeric JIS while recognizing the verified extra-mark defect for voiced katakana in version 1.12.57.
Verification status: See the evidence and test scope described in this guide. How formula examples are checked
JIS converts selected half-width characters to full-width equivalents. It can prepare ordinary ASCII letters, digits, and unvoiced half-width katakana for a width-specific display. Gnumeric 1.12.57 has a reproduced voiced-katakana limitation: JIS("ｶﾞ") produces ガ゛, including an extra spacing voiced mark. Do not use this version as an unattended general Japanese-text normalization step.
Syntax and supported scope
JIS(text)
text is the original string. The result is text with the implementation’s width mappings applied; other characters are copied. This is not a byte-count conversion. UTF-8 character widths in display and byte lengths are different concepts.
The name should not be interpreted as a complete conversion to a particular external legacy encoding. A workbook remains a Unicode-text workflow unless another export operation changes its encoding.
Worked examples verified in 1.12.57
| Formula | Native result |
|---|---|
=JIS("ABC123") |
ＡＢＣ１２３ |
=JIS("ｶﾀｶﾅ") |
カタカナ |
=JIS("A B") |
Ａ Ｂ |
=JIS("é") |
é |
=JIS("ｶﾞ") |
ガ゛ |
The ordinary space between A and B stays an ordinary space in the tested result. The accented Latin letter is copied unchanged. These examples deliberately limit what is certified: a few successful inputs do not establish correct conversion for every Japanese string or punctuation mark.
For a controlled ASCII identifier stored in A2, =JIS(A2) applies the same conversion pattern. Keep the original identifier in A2; full-width forms may not be valid identifiers in another system even when they look appropriate in a printed report.
Why the voiced-katakana case needs review
The source examines the next character while mapping the current katakana character. It can therefore form a voiced kana, but the main loop still processes the following half-width voiced mark, emitting a separate spacing mark. In the demonstrated input ｶﾞ, the output contains both ガ and ゛.
This is a bounded implementation observation, not a promise that all affected sequences have been enumerated. The corresponding conversion loop and relevant mapping behavior remain present in the separately retrieved 1.12.62 source, but no native 1.12.62 run is claimed.
Advanced use and safer alternatives
Use JIS only on a restricted, validated character repertoire whose output you have tested. For text containing voiced or semi-voiced half-width kana, retain the original and use a separately verified Japanese-text conversion process, or review and correct the affected output manually against the source text.
Do not blindly delete every voiced mark from the result: such marks can be meaningful, and a blanket replacement can corrupt valid data. Likewise, do not assume that composing JIS with another width function will repair the defect or round-trip arbitrary text exactly.
For matching and deduplication, decide whether width distinctions should be ignored at all. A display transformation is not automatically the canonical representation for names or identifiers.
Troubleshooting and version history
The official release notes add JIS in Gnumeric 1.9.10. The five outputs above were recalculated using the actual 1.12.57 engine under C.UTF-8 on 17 October 2026 with valid UTF-8 inputs only. Source for 1.12.62 was checked separately; source similarity is not a native result or a verified fix.
No invalid-UTF-8 tests, GUI rendering certification, or universal conversion guarantee is made. If an output has an extra mark, compare its characters with the original input rather than dismissing it as a font artifact. No retirement or last-supported version is established.