A lot of people here are clearly not reading the article.

It’s not about the <table> element itself—we hope everyone knows about tables—but rather about the table-specific DOM interface, including things like HTMLTableElement.prototype.insertRow() and HTMLTableRowElement.prototype.insertCell() as alternatives to the generic DOM techniques like Document.prototype.createElement() and Node.prototype.appendChild().

These are handy if you’re hand-writing your DOM interactions, but libraries that construct and maintain DOM trees (e.g. React, Svelte, Vue) will never use them, and that’s the direction everything has headed, so in practice they’ve fallen into near-complete disuse.

They match HTML syntax in another important way: HTML tables have thead/tbody/tfoot section elements, but you can mostly skip writing them in HTML syntax because it’ll imply <tbody> open and close tags. Likewise in this interface, if you have a thead/tbody/tfoot element you can call .insertRow() on it, but you can also call .insertRow() on the table, and it’ll put it in the last tbody, creating one if necessary. Meanwhile, I presume in React/Svelte/Vue/whatever you must write out your tbody manually.

I’ve definitely used at least .insertRow, .insertCell, .createTHead, .rows and .cells in the last five years in no-library throwaway/demo scripts where I was generating tables.

—⁂—

Concerning the specific example given, here’s what I find a clearer code style, using for instead of forEach, using better variable names, and omitting the index argument to insertRow/insertCell which was quite unnecessary and only confused matters (the author doesn’t seem to have realised it’s optional):

  let data = [
      ['one','two','three'],
      ['four','five','six']
  ];
  let table = document.createElement('table');
  for (const row of data) {
      let tr = table.insertRow();
      for (const value of row) {
          tr.insertCell().innerText = value;
      }
  }
  document.body.append(table);