HUB - Updated 2026-06-24
Excel CSV Encoding Fix: UTF-8 BOM, Shift-JIS, CP949
Fix garbled CSV files in Excel by comparing UTF-8, UTF-8 BOM, Shift-JIS, CP949/EUC-KR, delimiters, and upload-safe exports.
Open CSV Encoding FixerQuick Answer
Garbled CSV text usually means the file was opened with the wrong character encoding. Excel may guess UTF-8, UTF-8 with BOM, a local Windows code page, or Shift-JIS depending on the file and the user's environment. Keep the original file, inspect a small sample, compare multiple decodings, then choose the copy that matches the destination: Excel, web upload, API import, or a legacy system.
The repair order is important: make characters readable first, verify columns second, then export a destination-specific copy.
Start With the Symptom
Use this hub when the same CSV behaves differently in Excel, a browser upload, an API, or a Japanese/Korean legacy system.
| Search or Symptom | Open This First | Best Next Step |
|---|---|---|
excel character encoding | CSV Encoding Fixer | Compare UTF-8, UTF-8 BOM, Shift-JIS, and CP949 previews |
shift-jis to utf-8 | Shift-JIS to UTF-8 CSV | Confirm Japanese text before converting |
mojibake utf-8 shift_jis | Japanese CSV mojibake fix | Keep the original and export a clean copy |
| Excel is correct but upload fails | This guide | Try a BOM-free UTF-8 copy |
Why This Matters
CSV files move product catalogs, ad reports, customer lists, CRM exports, and logs between tools. When a file contains Korean, Japanese, accents, or emoji, a wrong encoding guess can turn names and addresses into unreadable text. The safest workflow is to separate the file source, the tool used to open it, and the system that will receive the final CSV.
Common Problems
| Symptom | Likely Cause | First Check |
|---|---|---|
| Korean text is garbled in Excel | UTF-8 read as a local Windows code page | Import as UTF-8 or use UTF-8 BOM |
| Japanese text shows mojibake | Shift-JIS and UTF-8 mismatch | Check whether the source is a Japanese Windows system |
| First column is strange after upload | UTF-8 BOM was not accepted | Test a BOM-free copy |
| Columns are shifted | Delimiter or quotes are wrong | Check comma, tab, and quoted fields |
| Preview score is high but names look wrong | Similar-looking legacy decoding | Compare real customer/product rows, not only headers |
Practical Workflow
- Save a backup of the original CSV.
- Test a few rows in CSV Encoding Fixer.
- Compare UTF-8, EUC-KR / CP949, and Shift-JIS candidate previews.
- Confirm the delimiter and column count after the characters are readable.
- If the target is Excel, try UTF-8 with BOM.
- If the target is an API or upload form, check whether BOM is allowed.
- For Japanese legacy files, verify Shift-JIS before converting to UTF-8.
How to Use the CSV Encoding Fixer
The tool is designed for a small, safe sample rather than a full private export.
| Step | What to Look For | Good Signal |
|---|---|---|
| Load a file | Candidate cards for UTF-8, EUC-KR / CP949, and Shift-JIS | One card has readable text and fewer mojibake signals |
| Check structure | Lines, columns, delimiter, and inconsistent rows | Header and sample rows keep the same number of columns |
| Select a decoding | Preview the actual Korean/Japanese rows | Names and addresses are readable, not just ASCII headers |
| Download a copy | Use UTF-8 BOM when the target is Excel | The downloaded copy reopens correctly in Excel |
Choose the Output by Destination
The same CSV may need different copies depending on where it goes next. Avoid trying to create one universal file when the audience is mixed.
| Destination | Recommended Copy | Check Before Sending |
|---|---|---|
| Excel review by a business user | UTF-8 with BOM or the recipient's expected local encoding | Open through Excel import and confirm headers plus one Japanese/Korean row |
| Web app upload | UTF-8, often without BOM | Confirm whether the first column name changes after upload |
| API ingestion | UTF-8 without Excel-specific changes | Check the API docs for charset and delimiter rules |
| Legacy Japanese system | Shift-JIS / CP932-style copy if required | Confirm unsupported characters before conversion |
| Mixed Korean/Japanese operations | UTF-8 for interchange, legacy copies only when required | Confirm both Korean and Japanese rows after conversion |
Triage Order for Garbled CSV
- Identify the source system before editing the file.
- Compare UTF-8, UTF-8 BOM behavior, EUC-KR / CP949, and Shift-JIS previews with the same two or three rows.
- Check delimiter and quote problems only after the characters are readable.
- Create a destination-specific copy instead of overwriting the original.
- Reopen the exported copy in the same tool the recipient will use.
Why UTF-8 Is Usually the Final Format
The modern web platform treats UTF-8 as the right interchange format for Unicode text, while legacy encodings remain necessary for deployed content and older systems. In practice, most cleaned CSV files should end as UTF-8, but the first diagnostic step may still need Shift-JIS or EUC-KR / CP949 so you can read the original bytes correctly.
For Excel handoff, UTF-8 with BOM can improve detection. For APIs and upload forms, plain UTF-8 is often safer because some systems treat the BOM as part of the first header.
Example
If a supplier sends a Japanese CSV and product names look like 譁ｰ讒倥し繧､繝医Ο繧ｰ, the file may be Shift-JIS but your tool is reading it as UTF-8. If a SaaS export is UTF-8 but Excel garbles it, a UTF-8 BOM copy may be better for Excel users.
Related Tool
- CSV Encoding Fixer: Compare encoding candidates, inspect delimiters, and create an Excel-friendly UTF-8 BOM copy.
Related Articles
Practical FAQ
Is CSV encoding the same as CSV delimiter?
No. Encoding decides how characters are decoded; delimiter decides where columns split. Fix garbled characters first, then inspect comma, tab, and quote issues.
Why does a CSV work in Excel but fail in an upload form?
Excel is tolerant and may guess a local encoding or accept a UTF-8 BOM. Upload forms are often stricter and may expect plain UTF-8, a specific delimiter, or no BOM in the first header.
Should the original CSV ever be overwritten?
No. Keep the original export unchanged, then create Excel-friendly or upload-friendly copies. This preserves the only reliable recovery point if a conversion is wrong.
Why does the tool show an encoding score instead of auto-fixing everything?
Encoding detection is not perfectly knowable from text alone. A score helps rank likely candidates, but a human still needs to confirm real data rows because headers may be ASCII-only and therefore look correct in several decodings.
When should delimiter problems be fixed?
After characters are readable. If Japanese or Korean text is still mojibake, delimiter fixes can hide the real problem. Once text is readable, compare comma, tab, semicolon, and quoted fields.
Updated
2026-06-24
Official references
These official references were used to verify the criteria discussed in this article.
- Import or export text (.txt or .csv) filesMicrosoft Support - Checked: 2026-06-24
- Encoding StandardWHATWG - Checked: 2026-06-24
- TextDecoderMDN Web Docs - Checked: 2026-06-24
- The Unicode StandardUnicode Consortium - Checked: 2026-06-24