CSV to Markdown table

The three things that break a hand-made Markdown table

• A pipe inside a cell. A value like b|c ends the cell early and shifts every column after it. Pipes are escaped here as \|.
• A newline inside a quoted CSV field. Markdown tables are one row per line, so an embedded newline splits the row. It is converted to <br>, which renders as a line break inside the cell on GitHub and most wikis.
• Ragged rows. A row with fewer fields than the header produces a malformed table. Short rows are padded with empty cells.

Aligned source, or compact source

With padding on, columns are padded with spaces so the raw Markdown lines up in a text editor — worth it for a table someone will edit by hand later. With it off you get the compact form, which is smaller in a diff and renders identically. Both are valid GitHub-flavored Markdown.

Turning off the header row

Markdown tables require a header row; there is no headerless table in the format. If your data has no header, uncheck the box and column letters (A, B, C) are used instead, so the table still renders and you can rename the columns afterward.

Where this fits

Documentation tables are usually generated from somewhere else — a query result, an export, a spreadsheet a colleague maintains. Converting once by hand is fine. Converting the same table on every release is a job for something that keeps the data.

TableDI is a desktop spreadsheet workspace that runs entirely on your machine. Its export can produce a REST endpoint for a table, which is a more durable way to keep a document in sync with data than pasting a table into it every time.