Free Markdown Table Generator

How the Markdown Is Generated

Markdown tables are not part of John Gruber's original 2004 specification. They arrived as an extension, standardised by GitHub Flavored Markdown and supported today by GitLab, most static site generators and nearly every documentation platform — but not by every parser. A minimal table needs exactly two things: a header line and a delimiter line beneath it.

| Product | Units | Revenue |
| ------- | ----: | ------: |
| Widget  |    12 |  120.00 |
| Gadget  |     4 |   99.50 |

The generator writes the header line first, then computes the delimiter for each column from that column's alignment setting and the width of its header text. Dashes are padded to the header's length so the source stays visually aligned, with a floor of three:

Body cells are then padded with trailing spaces to match the header width, so the pipes line up in a monospaced editor. That padding is purely cosmetic: Markdown ignores whitespace around cell contents, and a table written with no padding at all renders identically. It matters only to whoever has to read the raw file later — and in a README that people do edit by hand, it matters quite a lot.

Alignment is also a rendering instruction, not a data type. Setting a column to right-aligned does not tell the renderer the values are numbers; it just moves the text. Sorting, formatting and thousands separators are all beyond what Markdown tables can express.

What Fits in a Markdown Table Cell

Cells accept inline formatting and nothing more. That constraint is the source of nearly every problem people hit with Markdown tables, and knowing the boundary saves a lot of fighting with the syntax.

There is no way to put a line break inside a cell using Markdown alone. If a cell genuinely needs two lines, a <br> tag works on platforms that permit inline HTML, GitHub included. When several cells need that treatment, it is usually a sign the content wants a definition list or a set of subsections rather than a table.

The number of cells in each row should match the header. A row with too few cells renders with the missing columns blank on most parsers, which is untidy but survivable. A row with more cells than the header is the case to avoid: the surplus is discarded by renderers, and here it stops the Markdown output updating at all, so keep every row the same width as the header.

Importing CSV, and What Gets Lost

Import CSV reads a file from disk with the browser's FileReader API — the file is never uploaded. The first non-empty line becomes the header row and rebuilds the columns from scratch; every remaining line becomes a data row; blank lines are skipped. Each field is trimmed, and one pair of surrounding double quotes is removed.

The parser splits on every comma, which is where the limits appear. A field like "Smith, John" becomes two columns rather than one, because the quotes are only stripped after the split has already happened. Escaped double quotes inside a field ("" in CSV) are not unescaped, and a value containing a newline inside quotes is read as two separate rows. Files with CRLF endings are fine — the trailing carriage return is removed when each field is trimmed.

Importing also resets every column to left alignment, so set alignment afterwards rather than before. If your data contains commas, export it with a different delimiter and do a find-and-replace first, or paste the values into the grid by hand. For anything larger than a screenful, converting the CSV in a spreadsheet and pasting the result is usually faster than fixing a mangled import.

Use Cases

Related Tools