Benchmark of PowerShell de-duplication cmdlets

benchmark.ps1

#
# Benchmarking code used to produce results from blog post
# http://latkin.org/blog/2016/08/02/curious-behavior-when-de-duplicating-a-collection-in-powershell/
#

# home-grown HashSet-based approach
function hashunique {
  param(
      [Parameter(Mandatory=$true, ValueFromPipeline = $true)]
      [Object[]] $InputObject
  )
  begin   { $t = New-Object System.Collections.Generic.HashSet[Object] }
  process {
      foreach($x in $inputObject) {
          if($t.Add($x)) { $x }
        }
    }
}

1..20 |%{
    $collectionSize = $_ * 500
    $data = 1..$collectionSize |%{ Get-Random -Min 1 -Max ($collectionSize/2) }

    $selectTime = Measure-Command { $data | select -unique }
    $sortTime   = Measure-Command { $data | sort -unique }
    $groupTime  = Measure-Command { $data | group |% Name }
    $uniqTime   = Measure-Command { $data | hashunique }
    
    [PSCustomObject] @{
        'Collection Size' = $collectionSize
        'select -unique'  = $selectTime.TotalMilliseconds
        'sort -unique'    = $sortTime.TotalMilliseconds
        'group |% Name'   = $groupTime.TotalMilliseconds
        'hashunique'      = $uniqTime.TotalMilliseconds
    }
}