CODEWITHSUNDEEP
windowspowershellcmd
Laurie
Laurie

Reputation: 3

Create random data file from Windows CMD/PowerShell

I want to be able to generate a random data file in Windows using either CMD or PowerShell, with the generated data being comprised of lines and lines of random text. I've managed to achieve this in PowerShell using the following command, however this took around 1 minute to generate 1MB of data, which is way too slow to be generating GBs:

1..100000 | % { [System.Web.Security.Membership]::GeneratePassword(70, 3) >> C:/dummy.txt }

File output should be the following:

YyS@ZRU98udC3q#R@5o7AR$*Bh44v22J!ekKSpIAgLQyp^pbBx
s8Wm589aYYH39@Arb2^ZRMPjx2UaEwHYkMmhgFaU$QyAU@@@WU
yB^!qo6e4x*eFvx%ZY7738&&FkhHXU24OCJCxfyQ7a%peo!$ap
...........
...........
$GVhMrkZfJbIkgAgri0w9lFVt6a^vXh6ev&jwPHGfoE!pVW85r

Does anyone have any suggestions? Preferably I can do this without the need for external tools, as the script to create this data will be run automatically during startup of the machine.

Upvotes: 0

Views: 2747

Answers (2)

Mathias R. Jessen
Mathias R. Jessen

Reputation: 174515

The reason your current approach is slow is because of overhead from opening, writing to and closing the same file 100K times. Move the file redirection operator outside the pipeline expression and to only incur said overhead once:

1..100000 | % { [System.Web.Security.Membership]::GeneratePassword(70, 3) } >> .\dummy.txt

To illustrate the difference, here's measurements from 1000 passwords with your original vs. moving the output redirection outside:

PS ~>
>> Measure-Command {
>>   1..1000 | % { [System.Web.Security.Membership]::GeneratePassword(70, 3) >> .\dummy.txt }
>> } |Select TotalMilliseconds

TotalMilliseconds
-----------------
        8881.9736


PS ~>
>> Measure-Command {
>>   1..1000 | % { [System.Web.Security.Membership]::GeneratePassword(70, 3) } >> .\dummy.txt
>> } |Select TotalMilliseconds

TotalMilliseconds
-----------------
          72.7485

Literally >100 times faster already at only 1000 lines

Upvotes: 2

Rno
Rno

Reputation: 887

This should get you underway. The main point is that it uses C# for the generation of the random stuff. You can control the characters you want to use. It uses Linq, so it can be made to run much faster, but you should already see a massive performance boost when compared to random generation in PS. It will generate reasonably sized textfiles, if you want GB sized data you'll want to look into other methods.

$code = @"
using System;
using System.Linq;

namespace HelloWorld
{
    public class Program$id
    {
        const string chars = "abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789";
        private static Random random = new Random();
        
        public static string RandomString(int length) // infinitely faster than Get-Random in Powershell
        {
            return new string(Enumerable.Repeat(chars, length)
              .Select(s => s[random.Next(s.Length)]).ToArray());
        }

    }
}
"@

Add-Type -TypeDefinition $code -Language CSharp 
$lines = 100000
$lenght = 70
$outfile = "e:\temp\dummy.txt" # tweak as needed
$sb = [System.Text.StringBuilder]::new() # spin up a stringbuilder to hold characters
# create a string buffer with $lines of $length characters
1..$lines | % { 
    [void]$sb.AppendLine((Invoke-Expression "[HelloWorld.Program$id]::RandomString($lenght)"))
}
$sb.ToString() | Out-File $outfile # write out results
cat $outfile -Tail 10 # show last 10 lines in output file

This takes 2-3 seconds on my system (including the cat).

Upvotes: 1

Related Questions