CSV encoding and delimiter structure describe different parts of a text file. Encoding determines how bytes become characters. Delimiters separate fields. If an import shows strange letters or places every record in one column, those symptoms can require different investigations.

Changing commas to semicolons is unlikely to repair garbled names. Changing the character encoding is unlikely to fix a reader configured for the wrong separator.

Use the symptom to narrow the first check

If ordinary columns appear correctly but accented names are damaged, examine the encoding. If each line appears as one large cell while the letters look right, examine the delimiter and import settings.

These are clues, not definitive diagnoses. A file can have both problems, and quoted fields can contain separators or line breaks that a simplistic split misunderstands. Keep the original bytes available while investigating.

Text Encoding and Delimiter Inspector checks supported encoding evidence and quote-aware delimiter candidates. Its result can be inconclusive. Plain ASCII is compatible with several encodings, so the tool does not always have enough evidence to identify a unique source encoding.

A BOM is evidence, not an all-purpose repair

The W3C explanation of the byte-order mark describes how a marker can identify UTF-8. UTF-8 does not need a byte-order indicator for endianness, but some systems use the marker as an encoding signature.

An import application may have its own handling of that signature. Check its documented requirements instead of adding or removing a BOM solely because another file happened to work.

If there is no reliable encoding evidence, ask the source system how it exported the file. A guess that makes one name readable can still damage other characters. Avoid repeatedly saving a misread file, which can make the original bytes harder to recover.

Do not replace punctuation across the whole file

A sample field such as "North, West" contains a comma as part of its value. Globally replacing commas can change that content as well as the separators. Use a parser-aware conversion when changing the file's structure.

CSV Viewer can help inspect supported UTF-8 CSV or TSV. Check a row containing an accented name and another containing a quoted separator. Verify that both the characters and the field boundaries survive.

For an export handoff, record the encoding, delimiter and header expectation together. A note such as “UTF-8, comma-separated, one header row” gives the recipient a starting configuration. If a particular receiving system requires something else, follow that contract and test a representative sample through its preview.

A successful import should preserve the intended characters and columns. Seeing a table on screen is not enough if the values inside it have changed.