在 PHP 中，是否有一个函数返回一个由关联数组数组中的键值组成的数组?

所以，higher 很酷的事情-order 集合/迭代器函数，例如pluck, 过滤器，每个，map，和朋友们一起，他们可以混合搭配来组成更复杂的操作集.

So, the cool thing about higher-order collection/iterator functions such as pluck, filter, each, map, and friends is that they can be mixed and matched to compose a more complex set of operations.

大多数语言都提供这些类型的函数(寻找诸如集合、迭代器或枚举/枚举之类的包)……有些提供的函数比其他的多，您通常会看到这些函数在不同语言中的命名方式不同(即 collect == 地图，减少 == 折叠).如果函数在您的语言中不存在，您可以从存在的函数中创建它.

Most languages provide these types of functions (look for packages like collection, iterator, or enumeration/enumerable)...some provide more functions than others and you will commonly see that the functions are named differently across languages (i.e. collect == map, reduce == fold). If a function doesn't exist in your language, you can create it from the ones that do exist.

至于您的测试用例...我们可以使用 array_reduce 来实现 pluck.我发布的第一个版本依赖于array_map；但是，我同意 @salathe array_reduce 对这个任务更简洁；array_map 是一个不错的选择，但你最终不得不做更多的工作.array_reduce 乍一看可能有点奇怪，但如果回调井井有条，一切都会好起来的.

As for your test case...we can use array_reduce to implement pluck. The first version I posted relied on array_map; however, I agree with @salathe that array_reduce is more succinct for this task; array_map is an OK option, but you end up having to do more work in the end. array_reduce can look a bit odd at first, but if the callback is neatly organized, all is well.

一个不那么天真的 pluck 还会检查它是否可以在迭代值上调用"(一个函数/方法).在下面的简单实现中，我们假设结构是一个散列(关联数组).

A less naive pluck would also check to see if it can "call" (a function/method) on the iterated value. In the naive implementation below, we assume the structure to be a hash (associative array).

这将设置测试用例数据(Fixtures):

<?php

$data[] = array('categoryId' => 1,    'eventId' => 2,  'eventName' => 3,  'vendorName' => 4);
$data[] = array('categoryId' => 5,    'eventId' => 6,  'eventName' => 7,  'vendorName' => 8);
$data[] = array('categoryId' => 9,    'eventId' => 10, 'eventName' => 11, 'vendorName' => 12);
$data[] = array(/* no categoryId */   'eventId' => 10, 'eventName' => 11, 'vendorName' => 12);
$data[] = array('categoryId' => false,'eventId' => 10, 'eventName' => 11, 'vendorName' => 12);
$data[] = array('categoryId' => 0.0,  'eventId' => 10, 'eventName' => 11, 'vendorName' => 12);

选择您喜欢的 pluck 版本

$preferredPluck = 'pluck_array_reduce'; // or pluck_array_map

PHP 5.3+ 的pluck":array_reduce 提供了一个简洁的实现，虽然不像 array_map 版本那么容易推理:

function pluck_array_reduce($key, $data) {
  return array_reduce($data, function($result, $array) use($key){
    isset($array[$key]) &&
      $result[] = $array[$key];

    return $result;
  }, array());
}

PHP 5.3+ 的pluck":array_map 并不完美，所以我们必须做更多的检查(它仍然没有考虑到许多潜在的情况):

function pluck_array_map($key, $data) {
  $map = array_map(function($array) use($key){
    return isset($array[$key]) ? $array[$key] : null;
  }, $data);

  // is_scalar isn't perfect; to make this right for you, you may have to adjust
  return array_filter($map, 'is_scalar');
}

为旧版 PHP <5.3

我们可以使用遗留的create_function；然而，它的形式很糟糕，不推荐，也不优雅，因此，我决定不展示它.

We could have used the legacy create_function; however, it is bad form, not recommended, and also not at all elegant, thus, I've decided not to show it.

function pluck_compat($key, $data) {
  $map = array();
  foreach ($data as $array) {
    if (array_key_exists($key, $array)) {
      $map[] = $array[$key];
    }
  }
  unset($array);

  return $map;
}

这里我们根据我们运行的 PHP 版本选择一个版本的pluck"来调用.如果您运行整个脚本，无论您使用的是哪个版本，您都应该得到正确的答案.

$actual   = version_compare(PHP_VERSION, '5.3.0', '>=')
          ? $preferredPluck('categoryId', $data)
          : pluck_compat('categoryId', $data);
$expected = array(1, 5, 9, false, 0.0);
$variance = count(array_diff($expected, $actual));

var_dump($expected, $actual);
echo PHP_EOL;
echo 'variance: ', $variance, PHP_EOL;

print @assert($variance)
    ? 'Assertion Failed'
    : 'Assertion Passed';

注意没有结尾?>".那是因为它不需要.离开它比保留它更好.

Notice there is no ending '?>'. That is because it isn't needed. More good can come of leaving it off than from keeping it around.

FWIW，看起来这是作为 array_column.

FWIW, it looks like this is being added to PHP 5.5 as array_column.