AI says your CSV has four rows. Did it count lines instead of records?

A quoted CSV field can contain a line break. Count parsed records and state whether the header is included; do not split every visible line into a new record.

Python’s CSV documentation distinguishes the parser’s line count from returned records because a record can span multiple lines. Ask AI to identify the file’s delimiter and quoting rules before suggesting a cleanup.

Four physical lines, two data records

Our fictional CSV had a header and records A and B. A’s quoted note contained a line break and a comma; B’s note contained a quoted word. The text had four physical lines. Python’s CSV parser returned three rows including the header, meaning two data records.

Test punctuation inside the field

The parser preserved A’s multiline note and B’s embedded quote. Writing those parsed rows to an in-memory CSV and reading them again returned the same fields. Splitting the original by lines would produce four pieces; splitting by every comma would also break A’s note.

Ask for evidence before a rewrite

Suggested input: “Preserve each record ID. State delimiter, quote character and header handling. Show parsed record count and field count, including a sample with a quoted comma, embedded quote and line break. Do not delete lines to force the expected count.”

Keep the import assumptions visible

For Python file reading, the documentation uses newline handling appropriate to CSV, including newline set to an empty string. Other applications can use different dialects. Confirm the settings in the intended importer and check the actual fields after import.

Published October 14, 2026. Original fictional text was parsed and round-tripped in memory with existing Python 3.12.14. No AI service, spreadsheet import, customer export or general compatibility test was performed.

Sources: Python · CSV.