如何使用Powershell判断何时将一系列文件添加到网络

您可以使用以下批处理方法：

• 定义滑动时间窗口的长度，如果在其中创建了新文件，该时间窗口将重置；当事件到达该滑动窗口时，请继续收集事件。

  • 为防止集合无限期增长，请定义一个最大的批处理大小，即使要处理其他事件，也要处理该批处理。

• 在没有新事件到达的时间窗口过去之后，处理手头的批次，即到目前为止收集的事件，然后开始新的批次。

• 警告：

  • 该System.IO.FileSystemWatcher班可报告重复的事件。

  • 下面的代码消除了给定批次中的重复项，但消除了多个批次中的重复项，这将需要更多的精力-请参阅源代码注释。

• 实施说明：

  • 而不是使用-Action传递的脚本块Register-ObjectEvent来处理事件，而是在循环中对它们进行同步处理（带有超时）Wait-Event。

  • Wait-Event使用PowerShell的事件队列，因此通常不会错过事件（尽管在大容量情况下可能会在.NET级别发生）；相反，如果方法等待时碰巧到达，则FileSystemWatcher类似的WaitForChanged方法不会将事件排队，仅报告一个事件。

try {

  # Specify the target folder: the system's temp folder in this example.
  $dir = (Get-Item -EA Ignore temp:).FullName; if (-not $dir) { $dir = $env:TEMP }

  # Create and initialize the watcher.
  # Note the [ordered] to ensure that .EnableRaisingEvents is set last.
  $watcher = [System.IO.FileSystemWatcher] [ordered] @{
    Filter              = '*.datx'
    Path                = $dir
    EnableRaisingEvents = $true
  }

  # To simulate file creation, create *.datx files in the folder printed
  # mentioned in the following status message.
  Write-Host "Watching $dir for creation of $($watcher.Filter) files..."

  # Register for (subscribe to) creation events:
  # Determine a unique event-source ID...
  [string] $sourceId = New-Guid
  # ... and register for the watcher's Created event with it.
  Register-ObjectEvent $watcher -EventName Created -SourceIdentifier $sourceId

  # Initialize the ordered hashtable that collects all file names for a single 
  # batch.
  # Note: Since any given file creation can trigger *multiple* events, we 
  #       use an ordered hashtable (dictionary) to ensure that each file is 
  #       only reported once.
  #       However, *across batches* duplicates can still occur - see below.
  $batch = [ordered] @{}

  # Determine the sliding time window during which newly created files are 
  # considered part of a single batch.
  # That is, once a file has been created, each additional file created
  # within that time window relative to previous file becomes part of the
  # same batch.
  # When a time window elapses without a new file having been created, the
  # batch is considered complete and processed - see max. batch-size exception
  # below.
  # IMPORTANT:
  #  * The granularity is *seconds*, so the time window must be at least 1 sec.
  #  * Seemingly independently of the length of this window, duplicate events
  #    are likely to occur across batches the less time has elapsed between
  #    the end of a batch and the start of a new one - see below.
  $batchTimeWindowSecs = 5

  # How many names to allow a batch to contain at most, even if more
  # files keep getting created in the sliding time window.
  $maxBatchSize = 100

  while ($true) {
    # Run indefinitely; use Ctrl-C to exit.

    # Wait for events in a sliding time window of $batchTimeWindowSecs length.
    # Note: Using Wait-Event in a loop (1 event per iteration) is *more* 
    #       predictable than the multi-event collecting Get-Event in terms of
    #       avoiding duplicates, but duplicates do still occur.
    $batch.Clear()
    while ($evt = Wait-Event -SourceIdentifier $sourceId -Timeout $batchTimeWindowSecs) {
      $evt | Remove-Event  # By default, events linger in the queue; they must be removed manually.
      # Add the new file's name to the batch (unless already present)
      # IMPORTANT:
      #  * Duplicates can occur both in a single batch and across batches.
      #  * To truly weed out all duplicates, you'd have to create a session-level
      #    dictionary of the files' names and their creation timestamps.
      #    With high-volume file creation, this session-level dictionary could
      #    grow large; periodic removal of obsolete entries would help.
      $batch[$evt.SourceArgs.Name] = $null # dummy value; it is the *keys* that matter.
      Write-Host ✔ -NoNewline # status output: signal that a new file was created
      # If the max. batch size has been reached, submit the batch now, even if further
      # events are pending within the timeout window.
      if ($batch.Count -ge $maxBatchSize) { 
        Write-Warning "Max. batch size of $maxBatchSize reached; force-submitting batch."
        break
      }
    }

    # Completed batch available?
    if ($batch.Count) {      
      # Simulate processing the batch.
      Write-Host "`nBatch complete: Sending email for the following $($batch.Count) files:`n$($batch.Keys -join "`n")"                                            #`
      # Start a new batch.
      $batch.Clear()
    }
    else {
      Write-Host . -NoNewline # status output: signal that no new files were created in the most recent time window.
    }

  }

}
finally {
  # Clean up:
  # Unregister the event subscription.
  Unregister-Event -SourceIdentifier $sourceId
  # Dispose of the watcher.
  $watcher.Dispose() 
}

首先创建一批3个文件，然后再创建5个文件的示例输出：

Watching C:\Users\jdoe\AppData\Local\Temp for creation of *.datx files...
............✔✔✔
Batch complete: Sending email for the following 3 files:
1.datx
2.datx
3.datx
.✔✔✔✔✔
Batch complete: Sending email for the following 5 files:
4.datx
5.datx
6.datx
7.datx
8.datx
....................