在PHP累计计数性能下降 [英] Slow performance with cumulative count in PHP

问题描述

我有一个chartcontroller一些性能问题。我有两个数组：一个阵列使用日期和一个阵列与其中也包含一个日期的实体。我想累积计数的实体的日期值。

I have some performance issues in a chartcontroller. I have two arrays: one array with dates and one array with entities which contain also a date. I want to cumulative count the date values of the entities.

我确实有一点code，但它的速度慢（我必须等待3 - 第4秒钟的数据阵列中的68K行）。

I do have a bit of code but it's slow (I have to wait 3 - 4 seconds for 68k rows in the data array).

我的code：

// Loop through the SQL results and count on date
    foreach ($result as $row) {
        foreach ($dates as $date => $value) {
            // $row['appointmentDate'] is a DateTime object
            if ($date >= $row['appointmentDate']->format('Ymd')) {
                $dates[$date]++;
            }
        }
    }

有没有更聪明/更好/更快的方式来实现这一目标？

Is there a smarter / better / faster way to achieve this?

//修改

杰克红粉我，我需要推，我解决我的性能问题。

Jack gived me the push I needed and I solved my performance issue.

上面的code 38000了毫秒来完成，code 5700ms之下。

The code above took 38000 ms to complete, the code beneath 5700ms.

    usort($result, function($a, $b) {
        return $a['appointmentDate'] > $b['appointmentDate'];
    });

    // Loop through the SQL results and count on date
    for ($i = 0; $i++; $i <= count($result)) {
        $i++;
        $dates[$result[$i]['appointmentDate']->format('Ymd')] = $i;
    }

感谢您的帮助杰克！

Thanks for your help Jack!

推荐答案

是的。您可以通过做比较之前的两个数组排序实现了更好的性能。
随着两个数组排序就可以避免做的 N * M 循环，并保持在 N + M 的复杂性（加 N *日志N 和< STRONG> M * log M的<进行排序，这无论如何是小于N * M）/ STRONG>

Yes. You can achieve a better performance by sorting the two arrays before doing the comparison. With two arrays ordered you can avoid doing a n*m loop, and stay on n+m complexity (plus n*log n and m*log m for sorting, which anyway is less than n*m)

这篇关于在PHP累计计数性能下降的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！