Skip to content

Import from CSV or Excel

In this tutorial, you will let users drop or pick a CSV or Excel (.xlsx) file, parse it in the browser, and preview column headers before loading rows into Handsontable DataGrid. You will learn how to use PapaParse for CSV and the built-in ImportFile plugin for Excel, infer column types, and show clear errors.

Overview

This recipe shows a small UI with:

  • A drag-and-drop zone and a hidden file input.
  • File type detection by extension (.csv vs .xlsx) and routing to the right parser.
  • PapaParse for CSV and the built-in ImportFile plugin for Excel workbooks.
  • A header preview before you commit data to the grid.
  • Clear error messages for wrong type, empty files, and malformed content.

Difficulty: Intermediate
Time: ~20 minutes
Libraries: papaparse (npm in the docs build; a CDN script at runtime for your own HTML pages). Excel needs none: the ImportFile plugin is part of Handsontable.

Drop a .csv or .xlsx file here, or pick a source.

No data loaded yet

Drop a CSV or Excel file above, choose a file, or load the sample data to populate the table.

JavaScript
import Handsontable from 'handsontable/base';
import { registerAllModules } from 'handsontable/registry';
registerAllModules();
const CDN_PAPAPARSE = 'https://cdn.jsdelivr.net/npm/papaparse@5.7.0/papaparse.min.js';
const scriptPromises = new Map();
function loadScript(src) {
const cached = scriptPromises.get(src);
if (cached) {
return cached;
}
const p = new Promise((resolve, reject) => {
const existing = document.querySelector(`script[data-cdn="${src}"]`);
if (existing) {
existing.addEventListener('load', () => resolve(), { once: true });
existing.addEventListener('error', () => reject(new Error(`Failed to load ${src}`)), { once: true });
if (existing.getAttribute('data-loaded') === '1') {
resolve();
}
return;
}
const s = document.createElement('script');
s.src = src;
s.async = true;
s.dataset.cdn = src;
s.onload = () => {
s.setAttribute('data-loaded', '1');
resolve();
};
s.onerror = () => reject(new Error(`Failed to load ${src}`));
document.head.appendChild(s);
});
scriptPromises.set(src, p);
return p;
}
async function ensurePapa() {
if (typeof window.Papa !== 'undefined') {
return window.Papa;
}
await loadScript(CDN_PAPAPARSE);
if (typeof window.Papa === 'undefined') {
throw new Error('PapaParse did not register on window.');
}
return window.Papa;
}
function extensionOf(name) {
const i = name.lastIndexOf('.');
return i >= 0 ? name.slice(i + 1).toLowerCase() : '';
}
function showError(el, message) {
if (!el) {
return;
}
el.textContent = message;
el.hidden = false;
}
function clearError(el) {
if (!el) {
return;
}
el.textContent = '';
el.hidden = true;
}
function normalizeCellValue(value) {
if (value === null || value === undefined) {
return null;
}
if (typeof value === 'number' || typeof value === 'boolean') {
return value;
}
if (value instanceof Date) {
return value.toISOString().slice(0, 10);
}
const text = String(value).trim();
return text === '' ? null : text;
}
function mapRowByHeaders(row, headers) {
const out = {};
for (const key of headers) {
out[key] = normalizeCellValue(row[key]);
}
return out;
}
function processPapaResults(results) {
if (results.errors.length > 0) {
throw new Error(results.errors[0].message || 'CSV parse error.');
}
const fields = results.meta.fields?.filter((f) => f !== undefined && f !== '') ?? [];
if (fields.length === 0) {
throw new Error('No header row found in the CSV.');
}
const rows = results.data.map((row) => mapRowByHeaders(row, fields));
if (rows.length === 0) {
throw new Error('No data rows after the header.');
}
return { headers: fields, rows };
}
function parseCsvText(text, PapaRef) {
const trimmed = text.trim();
if (!trimmed) {
throw new Error('The file is empty.');
}
const parsed = PapaRef.parse(trimmed, {
header: true,
dynamicTyping: true,
skipEmptyLines: 'greedy',
transformHeader: (h) => h.trim(),
});
return processPapaResults(parsed);
}
async function parseCsvFile(file, PapaRef) {
return new Promise((resolve, reject) => {
PapaRef.parse(file, {
header: true,
dynamicTyping: true,
skipEmptyLines: 'greedy',
transformHeader: (h) => h.trim(),
complete: (results) => {
try {
resolve(processPapaResults(results));
} catch (e) {
reject(e instanceof Error ? e : new Error(String(e)));
}
},
error: (err) => reject(err instanceof Error ? err : new Error(String(err))),
});
});
}
async function parseXlsxFile(file) {
// A throwaway grid in a hidden element runs the ImportFile plugin. `apply: false` returns the
// mapped sheet without touching the grid, and a refused file rejects with a message that says why.
const host = document.createElement('div');
host.hidden = true;
document.body.appendChild(host);
const parser = new Handsontable(host, {
data: [[]],
importFile: true,
// licenseKey: 'your-license-key', // get a key at https://handsontable.com/pricing/
});
try {
const result = await parser.getPlugin('importFile').importFromBlob('xlsx', file, {
colHeaders: 'firstRow',
apply: false,
});
const rawHeaders = Array.isArray(result.colHeaders) ? result.colHeaders : [];
const headers = rawHeaders.map((cell, i) => {
const text = String(cell ?? '').trim();
return text === '' ? `Column ${i + 1}` : text;
});
if (headers.length === 0) {
throw new Error('No header row found in the Excel sheet.');
}
const rows = [];
for (const line of result.data) {
const obj = {};
let allEmpty = true;
headers.forEach((key, c) => {
obj[key] = normalizeCellValue(line[c]);
allEmpty = allEmpty && obj[key] === null;
});
if (!allEmpty) {
rows.push(obj);
}
}
if (rows.length === 0) {
throw new Error('No data rows after the header.');
}
return { headers, rows };
} finally {
parser.destroy();
host.remove();
}
}
async function parseFile(file) {
const ext = extensionOf(file.name);
if (ext === 'csv') {
const PapaRef = await ensurePapa();
return parseCsvFile(file, PapaRef);
}
if (ext === 'xlsx') {
return parseXlsxFile(file);
}
throw new Error('Unsupported file type. Use a .csv or .xlsx file.');
}
const SAMPLE_CSV = `Product,Category,In stock,Price
Widget A,Hardware,true,19.99
Widget B,Hardware,false,24.5
Service Pack,Services,true,0`;
function columnsFromHeaders(headers, rows) {
return headers.map((data) => {
const values = rows.map((row) => row[data]).filter((v) => v !== null);
if (values.length > 0 && values.every((v) => typeof v === 'number')) {
return { data, type: 'numeric' };
}
if (values.length > 0 && values.every((v) => typeof v === 'boolean')) {
return { data, type: 'checkbox' };
}
return { data, type: 'text' };
});
}
const gridContainer = document.querySelector('#example1');
const emptyEl = document.querySelector('#import-empty');
const errEl = document.querySelector('#import-error');
const fileInput = document.querySelector('#import-file');
const dropzone = document.querySelector('#import-dropzone');
const sampleBtn = document.querySelector('#import-load-sample');
let hot = null;
function loadIntoGrid({ headers, rows }) {
const columns = columnsFromHeaders(headers, rows);
if (!hot) {
if (emptyEl) {
emptyEl.hidden = true;
}
gridContainer.hidden = false;
hot = new Handsontable(gridContainer, {
data: rows,
columns,
colHeaders: headers,
rowHeaders: true,
height: 'auto',
width: '100%',
// licenseKey: 'your-license-key', // get a key at https://handsontable.com/pricing/
});
} else {
hot.updateSettings({ colHeaders: headers, columns });
hot.loadData(rows);
}
}
async function handleFile(file) {
clearError(errEl);
if (!file) {
return;
}
if (file.size === 0) {
showError(errEl, 'The file is empty.');
return;
}
try {
const payload = await parseFile(file);
loadIntoGrid(payload);
} catch (e) {
showError(errEl, e instanceof Error ? e.message : String(e));
}
}
fileInput?.addEventListener('change', () => {
const f = fileInput.files?.[0];
handleFile(f);
fileInput.value = '';
});
dropzone?.addEventListener('dragover', (ev) => {
ev.preventDefault();
dropzone.classList.add('import-dropzone--active');
});
dropzone?.addEventListener('dragleave', () => {
dropzone.classList.remove('import-dropzone--active');
});
dropzone?.addEventListener('drop', (ev) => {
ev.preventDefault();
dropzone.classList.remove('import-dropzone--active');
const f = ev.dataTransfer?.files?.[0];
handleFile(f);
});
sampleBtn?.addEventListener('click', async () => {
clearError(errEl);
try {
const PapaRef = await ensurePapa();
const payload = parseCsvText(SAMPLE_CSV, PapaRef);
loadIntoGrid(payload);
} catch (e) {
showError(errEl, e instanceof Error ? e.message : String(e));
}
});
TypeScript
import Handsontable from 'handsontable/base';
import { registerAllModules } from 'handsontable/registry';
import type { GridSettings } from 'handsontable/settings';
import Papa from 'papaparse';
registerAllModules();
type ParsedPayload = {
headers: string[];
rows: Record<string, string | number | boolean | null>[];
};
type ParsedCellValue = string | number | boolean | null;
type ParsedRow = Record<string, ParsedCellValue>;
const CDN_PAPAPARSE = 'https://cdn.jsdelivr.net/npm/papaparse@5.7.0/papaparse.min.js';
declare global {
interface Window {
Papa?: typeof Papa;
}
}
const scriptPromises = new Map<string, Promise<void>>();
function loadScript(src: string): Promise<void> {
const cached = scriptPromises.get(src);
if (cached) {
return cached;
}
const p = new Promise<void>((resolve, reject) => {
const existing = document.querySelector<HTMLScriptElement>(`script[data-cdn="${src}"]`);
if (existing) {
existing.addEventListener('load', () => resolve(), { once: true });
existing.addEventListener(
'error',
() => reject(new Error(`Failed to load ${src}`)),
{ once: true },
);
if (existing.getAttribute('data-loaded') === '1') {
resolve();
}
return;
}
const s = document.createElement('script');
s.src = src;
s.async = true;
s.dataset.cdn = src;
s.onload = () => {
s.setAttribute('data-loaded', '1');
resolve();
};
s.onerror = () => reject(new Error(`Failed to load ${src}`));
document.head.appendChild(s);
});
scriptPromises.set(src, p);
return p;
}
async function ensurePapa(): Promise<typeof Papa> {
if (typeof window.Papa !== 'undefined') {
return window.Papa;
}
await loadScript(CDN_PAPAPARSE);
if (typeof window.Papa === 'undefined') {
throw new Error('PapaParse did not register on window.');
}
return window.Papa;
}
function extensionOf(name: string): string {
const i = name.lastIndexOf('.');
return i >= 0 ? name.slice(i + 1).toLowerCase() : '';
}
function showError(el: HTMLElement | null, message: string): void {
if (!el) {
return;
}
el.textContent = message;
el.hidden = false;
}
function clearError(el: HTMLElement | null): void {
if (!el) {
return;
}
el.textContent = '';
el.hidden = true;
}
function normalizeCellValue(value: unknown): ParsedCellValue {
if (value === null || value === undefined) {
return null;
}
if (typeof value === 'number' || typeof value === 'boolean') {
return value;
}
if (value instanceof Date) {
return value.toISOString().slice(0, 10);
}
const text = String(value).trim();
return text === '' ? null : text;
}
function mapRowByHeaders(row: Record<string, unknown>, headers: string[]): ParsedRow {
const out: ParsedRow = {};
for (const key of headers) {
out[key] = normalizeCellValue(row[key]);
}
return out;
}
function processPapaResults(results: Papa.ParseResult<Record<string, unknown>>): ParsedPayload {
if (results.errors.length > 0) {
throw new Error(results.errors[0].message || 'CSV parse error.');
}
const fields = results.meta.fields?.filter((f) => f !== undefined && f !== '') ?? [];
if (fields.length === 0) {
throw new Error('No header row found in the CSV.');
}
const rows = (results.data as Record<string, unknown>[]).map((row) => mapRowByHeaders(row, fields));
if (rows.length === 0) {
throw new Error('No data rows after the header.');
}
return { headers: fields, rows };
}
function parseCsvText(text: string, PapaRef: typeof Papa): ParsedPayload {
const trimmed = text.trim();
if (!trimmed) {
throw new Error('The file is empty.');
}
const parsed = PapaRef.parse<Record<string, unknown>>(trimmed, {
header: true,
dynamicTyping: true,
skipEmptyLines: 'greedy',
transformHeader: (h) => h.trim(),
});
return processPapaResults(parsed);
}
async function parseCsvFile(file: File, PapaRef: typeof Papa): Promise<ParsedPayload> {
return new Promise((resolve, reject) => {
PapaRef.parse<Record<string, unknown>>(file, {
header: true,
dynamicTyping: true,
skipEmptyLines: 'greedy',
transformHeader: (h) => h.trim(),
complete: (results) => {
try {
resolve(processPapaResults(results));
} catch (e) {
reject(e instanceof Error ? e : new Error(String(e)));
}
},
error: (err) => reject(err instanceof Error ? err : new Error(String(err))),
});
});
}
async function parseXlsxFile(file: File): Promise<ParsedPayload> {
// A throwaway grid in a hidden element runs the ImportFile plugin. `apply: false` returns the
// mapped sheet without touching the grid, and a refused file rejects with a message that says why.
const host = document.createElement('div');
host.hidden = true;
document.body.appendChild(host);
const parser = new Handsontable(host, {
data: [[]],
importFile: true,
// licenseKey: 'your-license-key', // get a key at https://handsontable.com/pricing/
});
try {
const result = await parser.getPlugin('importFile').importFromBlob('xlsx', file, {
colHeaders: 'firstRow',
apply: false,
});
const rawHeaders = (Array.isArray(result.colHeaders) ? result.colHeaders : []) as unknown[];
const headers = rawHeaders.map((cell, i) => {
const text = String(cell ?? '').trim();
return text === '' ? `Column ${i + 1}` : text;
});
if (headers.length === 0) {
throw new Error('No header row found in the Excel sheet.');
}
const rows: ParsedRow[] = [];
for (const line of result.data as unknown[][]) {
const obj: ParsedRow = {};
let allEmpty = true;
headers.forEach((key, c) => {
obj[key] = normalizeCellValue(line[c]);
allEmpty = allEmpty && obj[key] === null;
});
if (!allEmpty) {
rows.push(obj);
}
}
if (rows.length === 0) {
throw new Error('No data rows after the header.');
}
return { headers, rows };
} finally {
parser.destroy();
host.remove();
}
}
async function parseFile(file: File): Promise<ParsedPayload> {
const ext = extensionOf(file.name);
if (ext === 'csv') {
const PapaRef = await ensurePapa();
return parseCsvFile(file, PapaRef);
}
if (ext === 'xlsx') {
return parseXlsxFile(file);
}
throw new Error('Unsupported file type. Use a .csv or .xlsx file.');
}
const SAMPLE_CSV = `Product,Category,In stock,Price
Widget A,Hardware,true,19.99
Widget B,Hardware,false,24.5
Service Pack,Services,true,0`;
function columnsFromHeaders(headers: string[], rows: ParsedRow[]): GridSettings['columns'] {
return headers.map((data) => {
const values = rows
.map((row) => row[data])
.filter((v): v is string | number | boolean => v !== null);
if (values.length > 0 && values.every((v) => typeof v === 'number')) {
return { data, type: 'numeric' as const };
}
if (values.length > 0 && values.every((v) => typeof v === 'boolean')) {
return { data, type: 'checkbox' as const };
}
return { data, type: 'text' as const };
});
}
const gridContainer = document.querySelector<HTMLElement>('#example1')!;
const emptyEl = document.querySelector<HTMLElement>('#import-empty');
const errEl = document.querySelector<HTMLElement>('#import-error');
const fileInput = document.querySelector<HTMLInputElement>('#import-file');
const dropzone = document.querySelector<HTMLElement>('#import-dropzone');
const sampleBtn = document.querySelector<HTMLButtonElement>('#import-load-sample');
let hot: Handsontable | null = null;
function loadIntoGrid({ headers, rows }: ParsedPayload): void {
const columns = columnsFromHeaders(headers, rows);
if (!hot) {
if (emptyEl) {
emptyEl.hidden = true;
}
gridContainer.hidden = false;
hot = new Handsontable(gridContainer, {
data: rows,
columns,
colHeaders: headers,
rowHeaders: true,
height: 'auto',
width: '100%',
// licenseKey: 'your-license-key', // get a key at https://handsontable.com/pricing/
});
} else {
hot.updateSettings({ colHeaders: headers, columns });
hot.loadData(rows);
}
}
async function handleFile(file: File | null | undefined): Promise<void> {
clearError(errEl);
if (!file) {
return;
}
if (file.size === 0) {
showError(errEl, 'The file is empty.');
return;
}
try {
const payload = await parseFile(file);
loadIntoGrid(payload);
} catch (e) {
showError(errEl, e instanceof Error ? e.message : String(e));
}
}
fileInput?.addEventListener('change', () => {
const f = fileInput.files?.[0];
handleFile(f);
fileInput.value = '';
});
dropzone?.addEventListener('dragover', (ev) => {
ev.preventDefault();
dropzone.classList.add('import-dropzone--active');
});
dropzone?.addEventListener('dragleave', () => {
dropzone.classList.remove('import-dropzone--active');
});
dropzone?.addEventListener('drop', (ev) => {
ev.preventDefault();
dropzone.classList.remove('import-dropzone--active');
const f = ev.dataTransfer?.files?.[0];
handleFile(f);
});
sampleBtn?.addEventListener('click', async () => {
clearError(errEl);
try {
const PapaRef = await ensurePapa();
const payload = parseCsvText(SAMPLE_CSV, PapaRef);
loadIntoGrid(payload);
} catch (e) {
showError(errEl, e instanceof Error ? e.message : String(e));
}
});
CSS
.import-csv-excel-wrap {
display: flex;
flex-direction: column;
margin: 0 -1rem;
}
.import-dropzone {
padding: 1.5rem 1rem;
text-align: center;
border: 0;
border-bottom: 1px solid var(--sl-color-gray-5, #e0e0e0);
border-radius: 0;
background: transparent;
color: var(--sl-color-gray-2, #555555);
font-size: var(--sl-text-xs, 0.75rem);
transition: background 0.15s ease;
}
.import-dropzone p {
margin: 0 0 0.75rem;
}
.import-dropzone.import-dropzone--active {
background: var(--sl-color-gray-6, #eeeeee);
}
.import-actions {
display: inline-flex;
align-items: center;
gap: 0.5rem;
}
.import-file-label {
display: inline-block;
cursor: pointer;
}
.import-file-label span,
.import-sample-btn {
display: inline-flex;
align-items: center;
padding: 0.375rem 0.75rem;
border: 1px solid var(--sl-color-gray-5, #e0e0e0);
border-radius: 0;
background: var(--sl-color-gray-6, #eeeeee);
color: var(--sl-color-gray-2, #555555);
font-family: var(--sl-font, Inter, system-ui, -apple-system, sans-serif);
font-size: var(--sl-text-xs, 0.75rem);
font-weight: 500;
line-height: 1.4;
cursor: pointer;
transition: background-color 0.15s, color 0.15s, border-color 0.15s;
}
.import-file-label:hover span,
.import-sample-btn:hover {
background: var(--sl-color-gray-5, #e0e0e0);
color: var(--sl-color-white, #333333);
}
.import-file-label input {
position: absolute;
width: 1px;
height: 1px;
opacity: 0;
pointer-events: none;
}
.import-msg {
margin: 0;
padding: 0.75rem 1rem;
border-radius: 0;
font-size: var(--sl-text-xs, 0.75rem);
}
.import-msg--error {
border-bottom: 1px solid var(--sl-color-red, #ef4444);
background: var(--sl-color-red-low, rgba(239, 68, 68, 0.12));
color: var(--sl-color-red, #ef4444);
}
.import-empty {
display: flex;
flex-direction: column;
align-items: center;
justify-content: center;
gap: 0.5rem;
padding: 2.5rem 1rem;
min-height: 200px;
color: var(--sl-color-gray-2, #555555);
text-align: center;
}
.import-empty[hidden] {
display: none;
}
.import-empty-icon {
display: inline-flex;
align-items: center;
justify-content: center;
width: 40px;
height: 40px;
margin-bottom: 0.25rem;
border: 1px solid var(--sl-color-gray-5, #e0e0e0);
color: var(--sl-color-gray-3, #777777);
}
.import-empty-title {
margin: 0;
color: var(--sl-color-white, #333333);
font-size: var(--sl-text-sm, 0.875rem);
font-weight: 600;
}
.import-empty-text {
margin: 0;
max-width: 38ch;
font-size: var(--sl-text-xs, 0.75rem);
line-height: 1.5;
}
#example1 {
padding: 1rem;
min-height: 180px;
}
#example1[hidden] {
display: none;
}
HTML
<div class="import-csv-excel-wrap">
<div
class="import-dropzone"
id="import-dropzone"
tabindex="0"
role="button"
aria-label="Drop a CSV or Excel file here"
>
<p>Drop a <code>.csv</code> or <code>.xlsx</code> file here, or pick a source.</p>
<div class="import-actions">
<label class="import-file-label">
<span>Choose file</span>
<input id="import-file" type="file" accept=".csv,.xlsx,application/vnd.openxmlformats-officedocument.spreadsheetml.sheet,text/csv" />
</label>
<button type="button" id="import-load-sample" class="import-sample-btn">Load sample data</button>
</div>
</div>
<div id="import-error" class="import-msg import-msg--error" hidden></div>
<div id="import-empty" class="import-empty">
<span class="import-empty-icon" aria-hidden="true">
<svg width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round">
<rect x="3" y="3" width="18" height="18"/>
<path d="M3 9h18M3 15h18M9 3v18M15 3v18"/>
</svg>
</span>
<p class="import-empty-title">No data loaded yet</p>
<p class="import-empty-text">Drop a CSV or Excel file above, choose a file, or load the sample data to populate the table.</p>
</div>
<div id="example1" hidden></div>
</div>

CDN scripts (no bundler)

For a plain HTML page, load Handsontable plus PapaParse from a CDN (pin the version to match what you test):

<script src="https://cdn.jsdelivr.net/npm/papaparse@5.7.0/papaparse.min.js"></script>

The JavaScript and React examples load the same URL on demand when window.Papa is missing, so the pattern matches a script-tag setup. The Angular and Vue examples import papaparse from npm instead.

  1. Accept files and detect type

    Use accept on the file input and check file.name to route .csv and .xlsx to the right parser. Reject everything else with a short message.

    The file input and drop zone (HTML)

    <div class="import-dropzone" id="import-dropzone" tabindex="0" role="button">
    <p>Drop a <code>.csv</code> or <code>.xlsx</code> file here, or pick a source.</p>
    <div class="import-actions">
    <label class="import-file-label">
    <span>Choose file</span>
    <input id="import-file" type="file"
    accept=".csv,.xlsx,application/vnd.openxmlformats-officedocument.spreadsheetml.sheet,text/csv" />
    </label>
    <button type="button" id="import-load-sample" class="import-sample-btn">Load sample data</button>
    </div>
    </div>

    What’s happening:

    • The accept attribute restricts the system file picker to .csv and .xlsx. This is a hint to the browser only — you must validate the extension in JavaScript as well.
    • The real <input> is visually hidden (a 1 px, absolutely positioned element with opacity: 0) and activated via the wrapping <label>. Clicking Choose file triggers the file picker without requiring a separate button.
    • Load sample data parses a bundled CSV string so users can try the grid without having to upload a file first.
    • The drop zone has tabindex="0" and role="button", so keyboard users can focus it. It has no key handler; keyboard users open the file picker with Choose file.

    Extension detection

    function extensionOf(name) {
    const i = name.lastIndexOf('.');
    return i >= 0 ? name.slice(i + 1).toLowerCase() : '';
    }

    What’s happening:

    • Takes the file name and returns the lowercase extension ("csv", "xlsx", or "" if there is none).
    • lastIndexOf('.') handles names like report.v2.csv correctly by taking the final dot.
    • The result is compared in parseFile and an error is thrown for anything that is not .csv or .xlsx.

    Wiring up file events

    fileInput?.addEventListener('change', () => {
    const f = fileInput.files?.[0];
    handleFile(f);
    fileInput.value = ''; // reset so the same file can be re-selected
    });
    dropzone?.addEventListener('dragover', (ev) => {
    ev.preventDefault();
    dropzone.classList.add('import-dropzone--active');
    });
    dropzone?.addEventListener('dragleave', () => {
    dropzone.classList.remove('import-dropzone--active');
    });
    dropzone?.addEventListener('drop', (ev) => {
    ev.preventDefault();
    dropzone.classList.remove('import-dropzone--active');
    const f = ev.dataTransfer?.files?.[0];
    handleFile(f);
    });

    What’s happening:

    • Both code paths (file input change and drop zone drop) hand the File object to the same handleFile function.
    • ev.preventDefault() on dragover is required — without it the browser opens the file instead of passing it to the drop handler.
    • Adding/removing import-dropzone--active gives visual feedback while a file is dragged over the zone.
    • Resetting fileInput.value = '' after processing allows the user to import the same file a second time.
  2. Parse CSV with PapaParse

    Lazy-load PapaParse from a CDN

    const CDN_PAPAPARSE = 'https://cdn.jsdelivr.net/npm/papaparse@5.7.0/papaparse.min.js';
    const scriptPromises = new Map();
    function loadScript(src) {
    const cached = scriptPromises.get(src);
    if (cached) return cached;
    const p = new Promise((resolve, reject) => {
    const existing = document.querySelector(`script[data-cdn="${src}"]`);
    if (existing) {
    existing.addEventListener('load', () => resolve(), { once: true });
    existing.addEventListener('error', () => reject(new Error(`Failed to load ${src}`)), { once: true });
    if (existing.getAttribute('data-loaded') === '1') resolve();
    return;
    }
    const s = document.createElement('script');
    s.src = src; s.async = true; s.dataset.cdn = src;
    s.onload = () => { s.setAttribute('data-loaded', '1'); resolve(); };
    s.onerror = () => reject(new Error(`Failed to load ${src}`));
    document.head.appendChild(s);
    });
    scriptPromises.set(src, p);
    return p;
    }
    async function ensurePapa() {
    if (typeof window.Papa !== 'undefined') return window.Papa;
    await loadScript(CDN_PAPAPARSE);
    if (typeof window.Papa === 'undefined') throw new Error('PapaParse did not register on window.');
    return window.Papa;
    }

    What’s happening:

    1. loadScript injects a <script> tag and returns a Promise that resolves on load and rejects on error.
    2. A Map keyed by URL caches the promise so the script is only injected once even if loadScript is called multiple times concurrently.
    3. It checks for a pre-existing <script data-cdn="..."> element — this handles the case where the tag was already added (e.g., in the docs preview runner).
    4. ensurePapa first checks window.Papa (set when your page loads PapaParse with a <script> tag), and only loads the CDN script when it is absent.

    Why lazy-load?

    • Avoids loading a parser the user might never need.
    • Works with any host page whether or not a CDN tag was already added — no duplicate downloads.

    Parse a CSV file

    async function parseCsvFile(file, PapaRef) {
    return new Promise((resolve, reject) => {
    PapaRef.parse(file, {
    header: true,
    dynamicTyping: true,
    skipEmptyLines: 'greedy',
    transformHeader: (h) => h.trim(),
    complete: (results) => {
    try {
    resolve(processPapaResults(results));
    } catch (e) {
    reject(e instanceof Error ? e : new Error(String(e)));
    }
    },
    error: (err) => reject(err instanceof Error ? err : new Error(String(err))),
    });
    });
    }

    What’s happening:

    • header: true tells PapaParse to treat the first row as column names and return each subsequent row as a plain object keyed by those names. results.meta.fields contains the ordered list of column names.
    • skipEmptyLines: 'greedy' drops rows that are entirely whitespace — handy for files with trailing newlines.
    • transformHeader: (h) => h.trim() removes accidental leading/trailing spaces from column names.
    • After parsing, empty or missing cell values are normalized to null so Handsontable displays a blank cell rather than the string "undefined" or "null".
    • dynamicTyping: true lets PapaParse return native numbers and booleans where possible.
    • Numeric and boolean values are preserved as-is; string values are trimmed, and empty strings are normalized to null.

    Parse a CSV text string (for the bundled sample)

    function parseCsvText(text, PapaRef) {
    const trimmed = text.trim();
    if (!trimmed) throw new Error('The file is empty.');
    const parsed = PapaRef.parse(trimmed, {
    header: true,
    dynamicTyping: true,
    skipEmptyLines: 'greedy',
    transformHeader: (h) => h.trim(),
    });
    return processPapaResults(parsed);
    }

    What’s happening:

    • Works identically to parseCsvFile but accepts a raw CSV string instead of a File object.
    • PapaParse’s synchronous parse(string, opts) overload is used here (no complete callback needed).
    • Used when the user clicks Load sample data to parse a CSV string embedded in the script and feed it into the grid.
  3. Parse Excel with the ImportFile plugin

    Read the workbook without a library

    The built-in ImportFile plugin reads .xlsx files, so Excel needs no parser of its own. The plugin runs on a grid, so the parser creates a throwaway one in a hidden element and destroys it afterwards:

    async function parseXlsxFile(file) {
    const host = document.createElement('div');
    host.hidden = true;
    document.body.appendChild(host);
    const parser = new Handsontable(host, {
    data: [[]],
    importFile: true,
    // licenseKey: 'your-license-key', // get a key at https://handsontable.com/pricing/
    });
    try {
    const result = await parser.getPlugin('importFile').importFromBlob('xlsx', file, {
    colHeaders: 'firstRow',
    apply: false,
    });
    const headers = (result.colHeaders ?? []).map((cell, i) => {
    const text = String(cell ?? '').trim();
    return text === '' ? `Column ${i + 1}` : text;
    });
    if (headers.length === 0) {
    throw new Error('No header row found in the Excel sheet.');
    }
    const rows = [];
    for (const line of result.data) {
    const obj = {};
    let allEmpty = true;
    headers.forEach((key, c) => {
    obj[key] = normalizeCellValue(line[c]);
    allEmpty = allEmpty && obj[key] === null;
    });
    if (!allEmpty) {
    rows.push(obj);
    }
    }
    if (rows.length === 0) {
    throw new Error('No data rows after the header.');
    }
    return { headers, rows };
    } finally {
    parser.destroy();
    host.remove();
    }
    }

    What’s happening:

    1. importFromBlob('xlsx', file, { colHeaders: 'firstRow', apply: false }) reads the first sheet and returns the mapped result without touching the grid: result.colHeaders holds the promoted header row and result.data the rows below it.
    2. A refused file (an .xls, a password-protected workbook, a damaged archive) rejects with the plugin’s own message, and handleFile shows it as it is. For an .xls or a password-protected file, the message says to save it as an unprotected .xlsx.
    3. Empty header cells get a fallback name (Column 1, Column 2, …), and rows that are entirely empty are skipped, so both parsers return the same { headers, rows } shape.
    4. normalizeCellValue keeps numbers and booleans, keeps the YYYY-MM-DD strings the plugin derives for date cells, trims other text, and turns blank cells into null.
    5. The throwaway grid is destroyed in finally, so a rejected file does not leak it. To import straight into a visible grid, with cell types, merges, and formulas (formulas need the Formulas plugin), call the same method on that grid without apply: false.

    Route to the right parser

    async function parseFile(file) {
    const ext = extensionOf(file.name);
    if (ext === 'csv') {
    const PapaRef = await ensurePapa();
    return parseCsvFile(file, PapaRef);
    }
    if (ext === 'xlsx') {
    return parseXlsxFile(file);
    }
    throw new Error('Unsupported file type. Use a .csv or .xlsx file.');
    }

    What’s happening:

    • extensionOf extracts the lowercase extension. Only .csv and .xlsx are handled; anything else throws immediately before any network request is made.
    • PapaParse is loaded on demand for CSV. Excel goes through the built-in plugin, so nothing is downloaded for it.
    • Both parsers return the same shape: { headers: string[], rows: object[] } — the rest of the code does not need to know which parser ran.
  4. Load parsed data into the grid

    Build column definitions from header names

    function columnsFromHeaders(headers, rows) {
    return headers.map((data) => {
    const values = rows
    .map((row) => row[data])
    .filter((v) => v !== null);
    if (values.length > 0 && values.every((v) => typeof v === 'number')) {
    return { data, type: 'numeric' };
    }
    if (values.length > 0 && values.every((v) => typeof v === 'boolean')) {
    return { data, type: 'checkbox' };
    }
    return { data, type: 'text' };
    });
    }

    What’s happening:

    • Handsontable’s columns option expects an array of column descriptors. Each descriptor’s data property is the key used to read from the row objects.
    • The helper infers numeric and checkbox columns when all non-empty values in that column are numbers or booleans. Mixed columns stay as 'text'.

    Lazy-init the grid on the first load

    let hot = null;
    function loadIntoGrid({ headers, rows }) {
    const columns = columnsFromHeaders(headers, rows);
    if (!hot) {
    emptyEl.hidden = true;
    gridContainer.hidden = false;
    hot = new Handsontable(gridContainer, {
    data: rows,
    columns,
    colHeaders: headers,
    rowHeaders: true,
    height: 'auto',
    width: '100%',
    // licenseKey: 'your-license-key', // get a key at https://handsontable.com/pricing/
    });
    return;
    }
    hot.updateSettings({ colHeaders: headers, columns });
    hot.loadData(rows);
    }

    What’s happening:

    • hot is created on the first successful import only. Before that, the page shows an empty-state panel (emptyEl) and gridContainer (the #example1 div) is kept hidden so users don’t see a single blank cell.
    • On the first call: hide the empty state, reveal the grid container, and instantiate Handsontable with the parsed data, columns, and headers.
    • On subsequent calls: reuse the existing instance — updateSettings({ colHeaders, columns }) reconfigures the grid’s column shape, then loadData(rows) replaces the data and triggers a full re-render. updateSettings must run before loadData so Handsontable knows which keys to read from the row objects.

    Load the bundled sample

    const SAMPLE_CSV = `Product,Category,In stock,Price
    Widget A,Hardware,true,19.99
    Widget B,Hardware,false,24.5
    Service Pack,Services,true,0`;
    sampleBtn?.addEventListener('click', async () => {
    clearError(errEl);
    try {
    const PapaRef = await ensurePapa();
    const payload = parseCsvText(SAMPLE_CSV, PapaRef);
    loadIntoGrid(payload);
    } catch (e) {
    showError(errEl, e instanceof Error ? e.message : String(e));
    }
    });

    What’s happening:

    • SAMPLE_CSV is a small CSV string embedded in the script — enough to demonstrate the grid without forcing the user to upload anything.
    • The handler ensures PapaParse is available, parses the sample synchronously, and hands the result straight to loadIntoGrid — no preview step in between.
    • File uploads use the same loadIntoGrid function (see handleFile in Step 5), so both code paths converge on a single grid-population helper.
  5. Handle errors

    Error display helpers

    function showError(el, message) {
    if (!el) return;
    el.textContent = message;
    el.hidden = false;
    }
    function clearError(el) {
    if (!el) return;
    el.textContent = '';
    el.hidden = true;
    }

    What’s happening:

    • Both functions guard against a null element (the container may be absent in some environments).
    • hidden is toggled so the error area takes no space when empty — no flickering or layout shift.
    • textContent is used to prevent XSS in error messages that might echo user-provided input.

    All error conditions covered

    ConditionMessage
    Extension is not .csv or .xlsx"Unsupported file type. Use a .csv or .xlsx file."
    File has zero bytes"The file is empty."
    CSV has parse errorsFirst error message from PapaParse
    CSV has no header row"No header row found in the CSV."
    CSV has no data rows"No data rows after the header."
    Excel workbook refused (an .xls, a password-protected or damaged file, a workbook past the import limits)The ImportFile plugin’s own message
    Excel sheet is empty"No header row found in the Excel sheet."
    Excel has no data rows"No data rows after the header."

    The handleFile async function is the single place where all parse errors are caught and forwarded to showError:

    async function handleFile(file) {
    clearError(errEl);
    if (!file) return;
    if (file.size === 0) {
    showError(errEl, 'The file is empty.');
    return;
    }
    try {
    const payload = await parseFile(file);
    loadIntoGrid(payload);
    } catch (e) {
    showError(errEl, e instanceof Error ? e.message : String(e));
    }
    }

    What’s happening:

    • file.size === 0 is checked before any async work to give an instant response for empty files.
    • Any error thrown by the parsers bubbles up here and is shown to the user.
    • On success, the parsed { headers, rows } payload goes straight to loadIntoGrid — no intermediate state to reset on failure.
    • The Load sample data handler uses the same try/catch shape, so sample-parse errors surface in the same error panel.

Try it quickly

Click Load sample data in the example to populate the grid from the bundled CSV string, or save the snippet from Step 4 as sample.csv and drop it on the zone.

How it works - complete flow

  1. User picks or drops a file (or clicks Load sample data) — the change, drop, or sample button event fires and reaches handleFile / the sample handler.
  2. File size check — zero-byte files are rejected immediately (file path only).
  3. Extension routing — parseFile reads the extension, sends .csv to PapaParse, and sends .xlsx to the ImportFile plugin. The sample path always uses PapaParse via parseCsvText.
  4. Parsing — CSV goes through PapaParse with header: true; Excel goes through the built-in ImportFile plugin with apply: false, which returns the promoted header row and the rows below it.
  5. Normalization — empty cells become null, and native numbers/booleans are preserved for CSV and Excel.
  6. Grid population — loadIntoGrid replaces colHeaders, columns, and the data. The first load swaps the empty-state panel for the grid.
  7. Grid renders — Handsontable re-renders with the new columns and data.
  8. Errors at any step — caught by handleFile or the sample-parse handler and shown in the error panel.

What you learned

  • How to detect the file type by extension and route CSV files through PapaParse and Excel files through the built-in ImportFile plugin with a single handler function.
  • How to ship a bundled CSV sample so users can demo the grid without uploading a file first.
  • How to replace the column configuration and the data on every import, and how to keep the grid behind an empty-state panel until the first import.
  • How to handle errors at each stage — file size, parsing, and grid load — and surface them in a dedicated error panel.

Next steps