PowerShell v3 Check File Headers

The following function is a very quick way to validate a set of file signatures. The files were provided to us in a project with a single, incorrect extension. Seeing as I need to process about 3.8 million files, there was no way I wanted to manually do that. So, I narrowed down the headers to the usual suspects for a project like this (.pdf and .tif) and wrote this function:

function Check-Header

{

       param(

             $path

       )

      

       # Hexidecimal signatures for expected files

       $pdf = 25504446;

       $TIFF_1 = 492049;

       $TIFF_2 = 49492A00;

       $TIFF_3 = 4D4D002A;

       $TIFF_4 = 4D4D002B;

            

       # Get content of each file (up to 4 bytes) for analysis

       ([Byte[]] $fileheader = Get-Content -Path $path -TotalCount 4 -Encoding Byte) |

       ForEach-Object {

             if(("{0:X}" -f $_).length -eq 1)

             {

                   $HeaderAsHexString += "0{0:X}" -f $_

             }

             else

             {

                   $HeaderAsHexString += "{0:X}" -f $_

             }

       }

      

       # Validate file header

       @($pdf, $tiff_1, $tiff_2, $tiff_3, $tiff_4) -contains $HeaderAsHexString

}

This function does a few things:

1. Takes a file path argument
2. Declares five known signatures (there are the headers we want files to have)
3. Reads the first 4 bytes of the file into a [Byte[]] array
4. Passes this byte array to a simple if/else statement to convert each byte from byte to a hexidecimal string
5. Compares an array of all known good signatures to see if any of them match the converted file signature 

If the -contains operator validates that one of the binary arrays matches our header the function returns true. If it does not find a match it returns false. On a directory of 1024 files this took just over 3.9 seconds on my test server. If I can get a straight run, I anticipate my 3.8 million file collection to take just a shade more than 4 hours. I will be doing some other manipulation, so, it will be considerably slower, but, in cases like this, it just goes to show there is no alternative to a good automated solution.