Importing XLSX

The XLSX adapter accepts Uint8Array bytes and returns a canonical workbook snapshot. It is synchronous and has no file-picker or filesystem dependency.

import { parseXlsx, createWorkbook } from "@byfungsi/nori";

// Browser boundary: `file` is a File selected by the host.
const imported = parseXlsx(new Uint8Array(await file.arrayBuffer()));
if (!imported.ok) {
  console.error(imported.error.reason, imported.error.message);
} else {
  // Retain these notes in your UI; successful import is not lossless fidelity.
  console.info(imported.value.warnings);
  const created = createWorkbook(imported.value.snapshot);
  if (!created.ok) console.error(created.error.message);
  else showWorkbook(created.value); // Your application's callback.
}

In Node, pass new Uint8Array(await readFile(path)) from node:fs/promises. A native application supplies bytes through its platform file adapter; Nori has no React Native binding yet.

Limits and errors

parseXlsx(bytes, options) accepts maxInputBytes (20,000,000 default), maxUncompressedBytes (100,000,000), and maxCells (200,000). Values must be positive safe integers. XlsxError.reason is invalid-file, unsupported, or limit.

Declared ZIP sizes limit ordinary input, not hostile resource consumption. For public uploads, the host owns upload policy and CPU/memory isolation. Worker scheduling belongs to a platform adapter; core does not assume Web Workers.

Layout and saved views

Multiple sheets, dimensions, merged regions, frozen panes, hidden rows/columns, worksheet value/custom filters and saved value-sort metadata are supported. Saved sorted rows remain in their existing order; edits do not automatically re-sort. Advanced and table-level filters/sorts emit warnings. See the precise support matrix.

The parser never exposes XML nodes, ZIP entries or OOXML names as canonical model fields. If you need to construct a workbook without XLSX, use parseWorkbookSnapshot or createEmptySnapshot.

Output and persistence

Use JSON.stringify(workbook.exportSnapshot()) to save canonical document data, and createWorkbook(JSON.parse(json)) to rehydrate with validation. There is no writeXlsx API yet. Formula caches in snapshots may differ from current calculated results; use getCellValue for display or downstream processing.