CSV VIEWER / COLUMN GUIDE

Why CSV columns look wrong

Three common problems come from reading separators and record endings as ordinary characters. Check the file’s structure before changing the data. CSV Viewer preserves the source while you reinterpret it.

1. The separator is a semicolon

Name;City
Ada;Paris
Lin;Rome

With comma selected, this is one column containing Name;City, Ada;Paris and Lin;Rome. Choose semicolon in File settings and use the first row as names. The result has two columns, Name and City, with two data rows: Ada / Paris and Lin / Rome.

Auto detection samples complete records. A sample with equally plausible comma and semicolon separators asks you to choose. For example, a,b;c can be read in either way; the viewer does not invent certainty.

2. A city contains a comma

Name,City
Ada,"New York, NY"
Lin,Paris

The quotes protect the comma in Ada’s city. The correct table has two data rows and two columns; Ada’s City value is the complete string New York, NY. Splitting each line on every comma would incorrectly make a third field. Keep comma selected and preserve the quotes.

A literal quotation mark inside a quoted field is doubled. "He said ""hello""" represents He said "hello". A quote inside an unquoted field or text following a closing quote produces an error rather than a guessed table.

3. A note contains a newline

Name,Note
Ada,"Line one
Line two"
Lin,Done

The correct result has two data rows, not three. Ada’s Note is one cell containing a newline. The newline inside quotes is part of a value; the newline after the closing quote ends the record. Inspect cell to see the complete note. A physical line count is therefore different from a CSV logical record count.

Headers, short rows and encoding

A row with too few fields stays in the table, with missing positions marked. An explicit empty field remains distinct from a missing field. A longer row creates additional columns. Duplicate and empty names receive positional display labels while the original names remain available for exports.

If the first record is actual data, turn off Use first row as column names. If characters look garbled, select the source encoding rather than changing the delimiter. The default accepts marked UTF-16 and valid UTF-8; legacy encodings require an explicit choice.

Use the Manual for all rules and export boundaries. CSV Viewer does not repair invalid data automatically. Keep the original file before making changes in another program.