Could not check your sign-in status.
JIM-JIN Data Care Lab
CSV Garbled Text Repair
Review unreadable characters and prepare a usable UTF-8 CSV.
Review repairs
Check that the candidate matches the original text. Successful UTF-8 conversion does not guarantee recovery of the original characters.
Compare the first 5 rows
Shows the first 8 columns of each row, up to 120 characters per cell. The entire file is processed when saving.
| Row | Appearance when opened as UTF-8 | Candidate to save |
|---|
Characters requiring review
Save as UTF-8
Saving stops if unreadable areas remain. Change the encoding or enter the affected text manually.
With a normal download, this page cannot confirm that saving has finished. Check the file at the saving destination.
How-to guide CSV Encoding Repair
Review unreadable text
and save your CSV as UTF-8.
Review garbled text in contact lists, product lists or CSV exports from older systems and save as UTF-8. For mixed encodings, compare candidates per row or cell during repair.
How to repair encoding
- STEP 01
Choose processing location and files
Normally keep local processing and choose or drop CSV, TSV or TXT files, up to 20 at once. Check delimiter and intended output use, then review repair details.
- STEP 02
Compare repair candidates
Compare the first five rows and text requiring review against the original content. Choose a candidate for saving or enter verified text manually if none is correct. Switch the reviewed file when several are selected.
- STEP 03
Confirm review and save each file
Check the box confirming you reviewed candidates and notes. For local processing, save as UTF-8. For server processing, create the UTF-8 file and then save it. If a save link appears, click it too.
Automatic detection is an estimate. Being convertible to UTF-8 does not mean the original text has been recovered. Review candidates before saving. Original files are not overwritten directly.
Local versus server processing
On your computer (default)
Reading, repair and exporting run in your browser. File contents are not sent to a server.
Closing or reloading the page loses the on-screen session. Save required results before finishing.
On the server
Files are temporarily stored in this site’s private storage. Upload starts when you select server processing and review repair details.
Inputs and results are automatically deleted within 60 minutes of acceptance. Actions and resuming do not extend the deadline. Save with time to spare.
- Resume server work
- If completed uploads remain within their deadline, reopen the page in the same browser and login state and resume review and saving. Review candidates again. If upload stopped partway through, remove that task and select files again.
- Remove work after saving
- Confirm that the file is saved on your device, then delete server work. For an earlier task, use its deletion action. Downloading alone does not delete server work.
Use local processing if server processing is unavailable. Both methods use the plan limits displayed above the tool.
Reviewing and correcting candidates
- Compare the first five rows
- Compare the UTF-8 interpretation and proposed saved text. The summary shows the first eight columns per row and 120 characters per cell. Omitted content is still processed in full when saving.
- Text requiring review
- Review 20 locations at a time and continue to the next 20. Compare row, column and candidates, then choose the saved interpretation. Locations without an explicit choice use the first candidate.
- Enter text manually
- Use this if no candidate is correct. Enter content verified against the original material. Changing candidates or text clears the confirmation box; review and confirm again.
- Unreadable text
- Saving stops while unresolved text remains. Enter verified text manually or change the source-encoding setting and review again.
- Rows with different column counts
- Rows differing from the first row’s column count are flagged. Columns are preserved rather than added or removed automatically. Check the delimiter and source data.
- Original CSV
- Row 2: 00123,東京 saved as UTF-8
Row 3: 00456,大阪 saved using a Japanese Windows encoding - Review candidates and save
- Both rows become UTF-8.
Confirm that 東京 and 大阪 are readable and leading zeros remain.
This is one supported type of mixed encoding. Data with encoding changes partway through a cell may not be repairable in the same way.
Input and output settings
- Delimiter
- Choose comma, tab or semicolon to match the source columns. Check this if data appears in one column. The selected delimiter is also used when saving.
- Source encoding
- Open detailed encoding settings. Automatic candidate detection, including mixed encodings, is normally suitable. It primarily checks UTF-8, CP932 and EUC-JP. Choose a known source encoding explicitly if available.
- UTF-16 files
- UTF-16 LE and BE are supported. A leading encoding marker allows automatic recognition; otherwise select the encoding. UTF-16 support assumes one encoding for the whole file.
- Open in Excel
- Choose this intended output use to save UTF-8 with a BOM encoding marker. This is the default.
- Import into another system
- Saves UTF-8 without a BOM. Follow the destination’s requirements.
- Output filenames
- Adds -utf8 to the original name. Tab-separated output uses .tsv; comma or semicolon output uses .csv. Multiple inputs are saved separately, not merged.
Numbers, date-like text and formulas are not automatically converted to numbers or calculation results. Excel or another app may interpret them differently according to its import settings.
Limits by plan
These are default limits. If settings changed, check the current limits above the tool.
- Guest
- 5 MB total, 50,000 rows total, 256 columns
- Free member
- 20 MB total, 300,000 rows total, 1,024 columns
- Pro member
- 80 MB total, 3,000,000 rows total, 16,384 columns
Size and rows are totals across all selected files. Columns are the maximum per row. Headers count as rows; quoted cell line breaks do not count as separate rows. Here, 1 MB equals 1,048,576 bytes.
Enterprise defaults match Pro. Select up to 20 files at once. Device storage or memory may prevent processing even within limits.
Frequently asked questions
Can one file with several encodings be repaired?
When UTF-8, CP932, EUC-JP or other encodings meet at row or cell boundaries, review candidates and standardize to UTF-8. Mid-cell encoding changes or lost original information may prevent complete repair.
Can I save immediately if no text is flagged?
Zero flagged locations does not guarantee correct detection. Review Japanese text and symbols in the comparison and inspect the saved file before using it.
Can I use Excel files?
.csv, .tsv and .txt are supported, not .xlsx or .xls. Export Excel to CSV or another supported format first. TXT is also read as table data using the selected delimiter.
Why did changing settings clear the results?
Changing encoding, delimiter or output settings requires review again. Individual selections and manual input must be repeated. If an earlier server task remains, remove it before processing with new settings.
Why cannot I find a file after saving?
If a save dialog opens, choose a destination and wait for writing to complete. Click any added save link. With ordinary downloads, this page cannot verify device save completion. Check browser downloads or the destination, then confirm the file was saved.
Does cancellation remove server files?
Server work may remain after stopping processing. Resume it or use its deletion action. If deletion fails, follow the guidance and retry. Unsaved results become eligible for automatic deletion at expiry.
Why does processing fail within size and row limits?
Additional limits apply: 8 MB per logical row including cell line breaks, 10,000 individual candidate choices per file, 64 KiB of manual text per cell and 320 MB output per file. Split files as needed. Structural CSV problems such as unclosed quotes must be corrected at the source.
Can I add rows or columns or find and replace?
This tool reviews encoding candidates and standardizes to UTF-8. It does not edit entire tables, add or remove rows or columns, perform find/replace or merge files.
