CSV characters garbled in Excel? Check encoding before editing
Quick answer: Preserve the untouched CSV and import it through Excel’s Data → From Text/CSV workflow instead of repeatedly saving the garbled result. Check readable characters, column separation and data types independently.
This guide focuses on desktop Excel for Windows with the Text/CSV import feature. It provides a diagnostic workflow, not a guarantee that every damaged source can be recovered.
Which kind of problem are you seeing?
| Symptom | Investigate first |
|---|---|
| Accented or non-Latin names appear as odd symbols | Text encoding and the source export. |
| A whole record appears in one column | The delimiter used to separate fields. |
| An identifier loses its first zero | How that column is interpreted as a data type. |
| A date becomes a different date | Date interpretation and the intended source format. |
More than one issue can occur in the same file. A readable name does not prove that an account code or date was imported correctly. Take one representative example from each important column before choosing an import method.
Keep a source you can restart from
Put the downloaded CSV in a folder named source-untouched. Work from a copy and record which system exported it. If a colleague sent the file, ask for its encoding, separator and date convention rather than guessing from how Excel first displayed it.
Consider a fictional practice record with a name, an identifier 00127, and a date intended to mean 6 May. Your acceptance check should preserve all three meanings. This example is a checklist design, not a tested import result.
Use Excel’s import route
Microsoft says a UTF-8 CSV saved with a BOM can normally open directly. Without that marker, its UTF-8 troubleshooting guide recommends the Power Query Text/CSV import route or the text import option.
- Open a blank workbook.
- Choose Data → From Text/CSV, or Get Data → From File → From Text/CSV, depending on your version.
- Select the untouched CSV copy and inspect the preview.
- Check names with accents or non-Latin characters against a known correct example. Use the source’s stated encoding when an encoding control is available.
- Choose the correct delimiter in the preview. Use Transform Data if columns need review before loading, or load only when the preview is acceptable.
Microsoft’s text import guide documents the preview, delimiter and loading options, and warns that default column interpretation can affect dates or leading zeros.
Verify meaning, not just appearance
- Compare a name containing a character that previously failed.
- Compare an identifier with leading zeros against the source record.
- Check an ambiguous date against an unambiguous known event.
- Compare the first and last expected records.
- Check the number of columns and whether a comma inside a description has displaced another field.
Write down each result. If characters are correct but identifiers are wrong, focus the next attempt on that column’s interpretation rather than changing the encoding again. This prevents a successful repair of one symptom from hiding another.
If the source already contains replacement characters
Ask the exporting system or sender for a fresh, correctly encoded export. Do not manually invent missing names from a partial display. A spreadsheet that looks tidy after editing can still contain guesses instead of the original records.
Save your reviewed working copy with a distinct filename. For an important handoff, ask the recipient to confirm a few representative records in their own workflow. Avoid replacing the only source file before that check.
Related: A simple file naming system · Check a restored copy before relying on it.