Handle commas and line breaks inside CSV fields

A physical line is not always a record

Test those real-world fields so the receiving application opens a usable table without another cleanup pass.

ParseMill exports reviewed invoices to Excel with separate summaries, line items, sources and review status. You can import an XLSX or CSV template and save a client setup. Run a real invoice through to the final workbook and check it in the application that receives your work.

Count logical records returned by the parser, not physical text lines. Keep expected column names and widths so an incorrect delimiter or quoting rule becomes visible before the data enters a spreadsheet.

Use this executed synthetic fixture

The actual fixture stores a newline between Line one and Line two. Its CSV writer quotes the comma-bearing description, the multiline note and the embedded quote, doubling the embedded quote inside the quoted field. The expected result is two data records plus one header, each with three fields.

RFC 4180: Common CSV format documents quoting for commas, double quotes and line breaks. Its common format is a useful baseline, while the recipient’s actual dialect still needs confirmation.

Reproduce the parser check

During preparation of this guide, Python 3.14.6 wrote the synthetic rows with csv.writer and read the resulting text with csv.reader. Exact list equality passed: three logical rows including the header, with three fields in each row. This verifies the local parser round trip; it does not claim that a spreadsheet application was tested.

1. Construct the expected rows as structured strings, including the real embedded newline and quote.
2. Write with a CSV library rather than string concatenation. Open text files with the library’s documented newline handling.
3. Read the saved file back with the same declared dialect and compare every field with the expected strings.
4. Import that file into the receiving spreadsheet and repeat the logical row, column and text checks. Record any application-specific changes.

Python: CSV reading and writing documents reader and writer behavior and newline handling. Use its quoting support instead of hand-escaping only the cases visible in one sample.

Catch common corruption patterns

An extra column after Valve usually indicates that the comma was treated as a delimiter inside the description. An extra record at Line two indicates a multiline-field problem. Missing quotation marks around wide indicate an escaping or display issue that needs inspection.

Do not fix these symptoms by deleting punctuation from the source text. Repair the serialization or import configuration and rerun the exact-string comparison. Preserve the original CSV while diagnosing the failure.

Verify the recipient’s result

Acceptance requires A1 and A2 to remain separate records with their original descriptions and notes. A visually wrapped cell is acceptable only if the stored field still contains the intended text and line break.

Include this fixture when evaluating ParseMill exports. It tests whether source descriptions survive the whole CSV handoff, beyond a successful extraction or file download.