Issue 629 - Use a generator for Cells::getAllCacheKeys to improve performance

[matt-allan]

Using a generator reduces memory usage and improves performance
when loading large spreadsheets.

This is:

- [x] a bugfix
- [ ] a new feature

Checklist:

• Changes are covered by unit tests
• Code style is respected
• Commit message explains why the change is made (see https://github.com/erlang/otp/wiki/Writing-good-commit-messages)
• CHANGELOG.md contains a short summary of the change
• Documentation is updated as necessary

Why this change is needed?

PHPSpreadsheet currently uses a lot of memory when loading large spreadsheets (see #648, #629). All of the coordinates in the spreadsheet are copied into a new array when the method Cells::getAllCacheKeys is called. Since the coordinates are concatenated with a new string this results in new strings being created in memory.

This result of this method is only ever passed to Psr\SimpleCache\Cacheinterface::getMultiple and Psr\SimpleCache\Cacheinterface::setMultiple. Since both of these methods accept an iterable we can use a generator instead. Using a generator means we don't need to build the entire array in memory and instead can return one value at a time as needed.

I benchmarked this with a ~100K row spreadsheet and saw a 16% improvement in run time and 20% improvement in memory usage. You can view a comparison here.

You can also view the individual profiles here:

• baseline profile
• with generator

This is the benchmark script I used:

<?php

require __DIR__ . '/vendor/autoload.php';

$reader = new \PhpOffice\PhpSpreadsheet\Reader\Xlsx();

$spreadsheet = $reader->load(__DIR__ . '/my_export_100k.xlsx');

I also attached the xlsx if you would like to run the benchmark yourself.

[PowerKiKi]

Excellent PR, thanks !