CSV Accents Look Broken in Excel: Import the File as UTF-8

Separate the file's bytes from the text you see

An encoding tells an application how to translate bytes into characters. The original UTF-8 bytes for an accented character can still be intact even when one application displays them incorrectly. That is why the first repair attempt should be a new import of the original file with the right decoding choice.

Our sample deliberately contains both Latin accents and Korean text. Inspect all four values, not just the ASCII header. A header such as city,contact can look correct under several encodings and therefore does not establish that the data rows were decoded correctly.

Choose UTF-8 in Excel's import preview

Open a blank workbook and use Data > Get Data > From File > From Text/CSV. Select the preserved CSV. In the preview, choose 65001: Unicode (UTF-8), or the equivalent UTF-8 option under File Origin. Wording and available import commands differ by Excel version.

Select the actual delimiter as a separate setting. For this sample it is a comma. Confirm the two columns and all four non-ASCII values before loading. If UTF-8 produces errors or replacement characters, do not force it: confirm the source encoding with the exporter.

1. Preserve the original CSV under a separate filename.
2. Import through Text/CSV and select the documented encoding.
3. Check München, José, 서울, and 김민수 in the preview.
4. Load only after both characters and field boundaries look correct.

Related help: Fix a separate problem: CSV data opening in one column

Understand what a BOM can and cannot do

A UTF-8 byte order mark is an optional initial signature with bytes EF BB BF. Microsoft documents normal opening of UTF-8 CSV files saved with a BOM. This can help Excel recognize a file when someone opens it directly, while an explicit import gives you a chance to inspect the encoding choice.

A BOM is not a different alphabet and does not repair damaged characters. Some downstream systems expect UTF-8 without it, so follow the receiver's file specification. Check the first header after import; a poorly handled signature should not become part of the field name.

é  → UTF-8 bytes C3 A9
ü  → UTF-8 bytes C3 BC
서울 → retained as 서울 after UTF-8 round-trip
김민수 → retained as 김민수 after UTF-8 round-trip
Optional file-start BOM → EF BB BF

Export a corrected copy and inspect it again

Once the workbook contains the correct characters, save the working workbook if you need its spreadsheet features. When a CSV is required, choose a UTF-8 CSV export option where available, save to a new filename, and check that the intended sheet was exported.

Reimport that new file with the recipient's expected settings and compare the sample names, header count, and row count. If the characters were already wrong inside the workbook, correct them from a trusted source before exporting. Encoding the string José as UTF-8 simply preserves the wrong string in a different byte representation.

Recognize the limits of character recovery

If a prior export replaced unsupported characters with literal question marks, those characters may no longer be recoverable from that file alone. Look for the original export or source records. Avoid bulk substitutions that guess which names a corrupted string was meant to contain.

A UTF-8 file can also contain valid but already garbled text. Passing an encoding check therefore does not prove linguistic correctness. The Column Harbor checker reads UTF-8 CSV and flags selected risks; it does not transcode legacy encodings or restore damaged names. Compare important values with the original records after any cleanup.