Stream API

XLSX.stream creates Node Readable streams from an already loaded worksheet. It reduces the need to hold a complete output string, but does not stream input parsing or avoid storing the worksheet. The namespace is absent from the mini distribution.

Signatures

Function	Parameters	Returns
`to_csv(worksheet, options?)`	Worksheet and CSV conversion options	Text-output `Readable`
`to_html(worksheet, options?)`	Worksheet and HTML conversion options	Text-output `Readable`
`to_json(worksheet, options?)`	Worksheet and JSON conversion options	Object-mode `Readable`, one row per chunk
`set_readable(Readable)`	A compatible `Readable` constructor	No result to consume

Node registers its Readable constructor automatically. Use set_readable only when supplying that constructor yourself.

CSV

CSV options match utils.sheet_to_csv:

Option	Meaning
`FS`	Field separator, default comma
`RS`	Record separator, default newline
`strip`	Remove trailing separators
`blankrows`	Keep blank rows, default `true`
`skipHidden`	Skip hidden rows and columns
`forceQuotes`	Quote every field
`rawNumbers`	Use raw numeric values
`dateNF`	Date format override

CSV streams begin with a Unicode byte-order mark (\uFEFF). The ordinary sheet_to_csv string does not include this mark.

Consume a stream with a pipe, pipeline, or async iteration. Catch stream errors through the consumption method, not just a try around construction.

const XLSX = require("@agent-sheet/wasm");
const assert = require("node:assert/strict");
(async () => {
  const sheet = XLSX.utils.aoa_to_sheet([["Item", "Qty"], ["Pens", 3]]);
  let output = "";
  for await (const chunk of XLSX.stream.to_csv(sheet)) output += chunk.toString();
  assert.equal(output, "\uFEFF" + XLSX.utils.sheet_to_csv(sheet));
})().catch((error) => { throw error; });

HTML

Options are id, editable, header, footer, and gridcolor, matching utils.sheet_to_html. Without !ref, a range is inferred from existing cells; an empty worksheet produces a table with a blank A1 cell. The original raises an unhandled stream error that can terminate the process. See Known differences.

const XLSX = require("@agent-sheet/wasm");
const assert = require("node:assert/strict");
(async () => {
  const sheet = XLSX.utils.aoa_to_sheet([["Hello"]]);
  let html = "";
  for await (const chunk of XLSX.stream.to_html(sheet)) html += chunk.toString();
  assert.ok(html.includes("Hello"));
  assert.ok(html.includes("<table"));
})().catch((error) => { throw error; });

JSON rows

to_json returns row objects, not a serialized JSON array. Use header: 1 for array rows, header: "A" for column-letter keys, or an explicit header array for chosen keys. Other options include range, defval, raw, rawNumbers, blankrows, dateNF, skipHidden, and UTC.

const XLSX = require("@agent-sheet/wasm");
const assert = require("node:assert/strict");
(async () => {
  const sheet = XLSX.utils.aoa_to_sheet([["Name", "Score"], ["Ari", 8], ["Bo", 6]]);
  const rows = [];
  for await (const row of XLSX.stream.to_json(sheet)) rows.push(row);
  assert.deepEqual(rows, [{ Name: "Ari", Score: 8 }, { Name: "Bo", Score: 6 }]);
})().catch((error) => { throw error; });

Do not pipe an object-mode JSON stream directly to a byte-oriented file stream. Transform each row into text first, or consume the objects in your application.

See Streams for file output and Converting sheets for conversion behavior.