Reading Workbooks

Use readXlsx when you already have workbook bytes in memory.

Read from a buffer

import { readFile } from "node:fs/promises";
import { readXlsx } from "tablsx";

const bytes = await readFile("report.xlsx");
const workbook = readXlsx(bytes);

console.log(workbook.sheets.map((sheet) => sheet.name));

Use WorkbookReader for convenient access

import { readFile } from "node:fs/promises";
import { WorkbookReader } from "tablsx";

const bytes = await readFile("report.xlsx");
const reader = WorkbookReader.fromBuffer(bytes);

const sheet = reader.sheet("Orders");
console.log(sheet.rowCount);
console.log(sheet.getCell(1, 0).value);

Convert to objects

If the first row contains headers, SheetReader#toObjects() is the fastest path into application data:

const rows = reader.sheet("Orders").toObjects();

If you need explicit typing for a column, pass overrides:

import { CellType } from "tablsx";

const rows = reader.sheet("Embeddings").toObjects({
  columns: {
    embedding: { type: CellType.VECTOR },
    created_at: { type: CellType.DATE },
  },
});

Error behavior

readXlsx throws when required package parts are missing or inconsistent. That is intentional. tablsx is designed to fail fast on malformed workbook structure instead of silently guessing.