Powershell: Measure folder traversal methods

powershell-measure-folder-traversal-methods.ps1

param(
     [Parameter(Mandatory=$true)][string]$Filter,
     [Parameter(Mandatory=$true)][string]$Path
)

$current_year = (get-date).year;
$logging_time_format = "yyyyMMdd\tHHmmss";
$logging_time = Get-Date -Format $logging_time_format;
write-output "${logging_time}: Checking $Path for files that match filter ${Filter}.";
if (!(test-path $Path)) {
    $logging_time = Get-Date -Format $logging_time_format;
    write-output "${logging_time}: Aborting. No such folder: $Path.";
    exit -1;
}

echo "Starting test A";
$subjects_A = @();
$measure_A = Measure-Command -Expression {
  [reflection.assembly]::loadwithpartialname("Microsoft.VisualBasic") | Out-Null
  $xs = @();
  $date = (get-date).AddDays(-7);
  try {
    $xs = [Microsoft.VisualBasic.FileIO.FileSystem]::GetFiles(
        $Path,
        [Microsoft.VisualBasic.FileIO.SearchOption]::SearchTopLevelOnly,
        $Filter
    );
  } catch { $_ }
  $subjects_A = $xs | where {$_.LastWriteTime -le $date} `
                    | select -last 500 ;
}
echo "Done with test A. Results in ms: ";
$measure_A.TotalMilliseconds;

echo "Starting test B.";
$subjects_B = @();
$measure_B = Measure-Command -Expression {
  $date = (get-date).AddDays(-7);
  $subjects_B = [System.IO.Directory]::EnumerateFiles(
        $Path,
        $Filter,
        "TopDirectoryOnly"
    ) | where {[System.IO.File]::GetLastWriteTime($_) -le $date} `
      | select -last 500;
}
echo "Done with test B. Results in ms: ";
$measure_B.TotalMilliseconds;


echo "Starting test C.";
$subjects_C = @();
$measure_C = Measure-Command -Expression {
  $date = (get-date).AddDays(-7);
  $subjects_C = cmd.exe /c dir /B /O:D $Path\$Filter `
  | select -first 500 `
  | % { Join-Path -Path $Path -ChildPath $_} `
  | % { [System.IO.FileInfo]($_) } `
  | where {$_.LastWriteTime -le $date} | % { $_.FullName };
}
echo "Done with test C. Results in ms: ";
$measure_C.TotalMilliseconds;

echo "Starting test D.";
$subjects_D = @();
$measure_D = Measure-Command -Expression {
  $date = (get-date).AddDays(-7);
  $subjects_D = cmd.exe /c dir /B /O:D $Path\$Filter `
  | select -first 500 `
  | % { Join-Path -Path $Path -ChildPath $_} `
  | where { [System.IO.File]::GetLastWriteTime($_) -le $date};
}
echo "Done with test D. Results in ms: ";
$measure_D.TotalMilliseconds;


echo "Starting test E.";
$subjects_E = @();
$measure_E = Measure-Command -Expression {
    $date = (get-date).AddDays(-7);
    $subjects_E = cmd.exe /c dir /B /O:D $Path\$Filter `
    | % { Join-Path -Path $Path -ChildPath $_} `
    | %{ [pscustomobject]@{
        "date" = [System.IO.File]::GetLastWriteTime($_)
        "file" = "$($Path)\$_"
        }
    } `
    | where { $_.date -lt $date } `
    | select -last 500 `
    | % { $_.file };
}
echo "Done with test E. Results in ms: ";
$measure_E.TotalMilliseconds;

<# 

echo "";
$len = $subjects_A.Count;
echo "Test A found $len files: ";
$subjects_A;

echo "";
$len = $subjects_B.Count;
echo "Test B found $len files: ";
$subjects_B;

echo "";
$len = $subjects_C.Count;
echo "Test C found $len files: ";
$subjects_C;

echo "";
$len = $subjects_D.Count;
echo "Test D found $len files: ";
$subjects_D;

echo "";
$len = $subjects_E.Count;
echo "Test E found $len files: ";
$subjects_E;

#>

Use this Powershell script to test which method of traversing folder contents works fastest in your Windows environment.

It runs 5 tests, each slightly different to a lot different.

Test A is the most different as it uses a Visual Basic file enumeration component. In all tests conducted thus far, test A is significantly faster (about 10x) than the others.