How the CSV is read
CSV looks simple until a field contains a comma. The parser follows RFC 4180, the informational spec most tools export to, and reads it leniently:
- Fields in double quotes can contain commas, line breaks and doubled quotes (
""). - Rows can end in CRLF (Windows), LF (Mac and Linux) or a lone CR (old Mac exports).
- A byte order mark at the start (common in Excel’s “CSV UTF-8”) is removed.
- Blank lines are skipped, and a trailing newline doesn’t add an empty row.
- The delimiter is detected from the first rows: whichever of comma, semicolon, tab or pipe splits them into the most consistent number of columns. European exports often use semicolons because the comma is the decimal mark.
If a quote is never closed, everything after it lands in one cell. The tool warns you with the line where the quote opened.
What the table looks like
The output is a GitHub Flavored Markdown table. Tables are a GFM extension, so they render on GitHub, GitLab, most docs sites and MDX blogs, but not in a strict CommonMark renderer. A pipe in your data becomes \| and a line break becomes <br>, since a table row has to stay on one line.
Number columns can be right-aligned automatically (a heuristic: every non-empty cell looks like a number, currency or percentage), or you can set each column by hand. “Pad columns” lines the raw text up, which makes diffs easier to read. It counts characters, so emoji and wide CJK text can still look uneven in the source; the rendered table is fine.
Other data
For API responses use JSON to Markdown, and for feeds or sitemaps XML to Markdown. If the table lives on a web page, copy the page’s HTML into HTML to Markdown. Writing a post around it? The word counter and reading time calculator help with length.