Large files

Fast PHP Excel Library for Large Files

Use MNB PHPExcel as a fast, lightweight and memory-efficient PHP XLSX reader with forward-only rows, stream chunk processing and selective column reads.

Updated July 30, 2026

PHP 8.1+Individual XLSXStreamingLow memory

LevelIntermediateReading time14 minPackagemnb/mnb-phpexcel

MnbExcel::read()MnbExcel::largeRead()ReadSession::rows()LargeXlsxReadSession::chunk()ReadSession::selectColumns()

What you will learn

Choose between the full package and the individual XLSX package.
Stream worksheet rows without collecting the complete workbook into a PHP array.
Process predictable chunks and measure peak memory with your own production files.

Choose the package entry point#

composer require mnb/mnb-phpexcel:^2.0

composer require mnb/mnb-phpexcel-xlsx:^2.0

The full package provides the MnbExcel facade and large-file helpers. The individual XLSX package keeps the install focused and uses Mnb\PHPExcel\Format\Xlsx.

Stream rows instead of building one large array#

<?php

use Mnb\PHPExcel\MnbExcel;

$sheet = MnbExcel::read(__DIR__ . '/orders.xlsx')
    ->sheetOrActive('Orders')
    ->withHeaderRow(1)
    ->withOptions([
        'skip_empty_rows' => true,
        'include_cell_metadata' => false,
    ])
    ->selectColumns(['Order ID', 'Customer', 'Amount']);

foreach ($sheet->rows() as $row) {
    processOrder($row);
}

<?php

use Mnb\PHPExcel\Format\Xlsx;

$sheet = Xlsx::read(__DIR__ . '/orders.xlsx')
    ->sheetOrActive('Orders')
    ->withHeaderRow(1)
    ->withOptions([
        'skip_empty_rows' => true,
        'include_cell_metadata' => false,
    ])
    ->selectColumns(['Order ID', 'Customer', 'Amount']);

foreach ($sheet->rows() as $row) {
    processOrder($row);
}

Use the PHP Excel stream chunk reader#

<?php

use Mnb\PHPExcel\MnbExcel;

MnbExcel::largeRead(__DIR__ . '/orders.xlsx')
    ->sheet('Orders')
    ->withHeader()
    ->onlyColumns(['Order ID', 'Customer', 'Amount'])
    ->memoryGuardRatio(0.80)
    ->chunk(1000, function (array $rows, array $state): void {
        importOrderBatch($rows);
        printf("Processed %d rows\n", $state['rows_read'] ?? 0);
    });

<?php

use Mnb\PHPExcel\Format\Xlsx;

$rows = Xlsx::read(__DIR__ . '/orders.xlsx')
    ->sheetOrActive('Orders')
    ->withHeaderRow(1)
    ->withOptions(['skip_empty_rows' => true])
    ->rows();

$batch = [];
foreach ($rows as $row) {
    $batch[] = $row;

    if (count($batch) === 1000) {
        importOrderBatch($batch);
        $batch = [];
    }
}

if ($batch !== []) {
    importOrderBatch($batch);
}

Measure peak memory and elapsed time#

<?php

$startedAt = hrtime(true);
$startingMemory = memory_get_usage(true);
$rowCount = 0;

foreach ($sheet->rows() as $row) {
    ++$rowCount;
    processOrder($row);
}

$result = [
    'rows' => $rowCount,
    'seconds' => round((hrtime(true) - $startedAt) / 1_000_000_000, 3),
    'memory_delta_mb' => round((memory_get_usage(true) - $startingMemory) / 1048576, 2),
    'peak_memory_mb' => round(memory_get_peak_usage(true) / 1048576, 2),
];

print_r($result);

Run each benchmark in a fresh PHP process. PHP peak-memory counters are process-wide, so comparing multiple libraries inside one process can produce misleading results.

Keep the large-file path lightweight#

Use rows() or largeRead() instead of toArray() for unbounded files.
Select only the columns required by the import or report.
Disable cell metadata when values are sufficient.
Install ext-zip and ext-xmlreader for the strongest streaming path.
Process and persist each batch before requesting the next batch.
Benchmark with representative formulas, shared strings, dates and row widths—not only row count.

For a fair side-by-side test, continue to MNB PHPExcel vs PhpSpreadsheet performance benchmark.

Was this guide useful?Use GitHub issues for corrections, missing examples, or unclear behavior.

Open an issue