How to validate dependent parameters in PowerShell cmdlet

Question

What's the best way to do PowerShell cmdlet validation on dependent parameters? For example, in the sample cmdlet below I need to run validation that Low is greater than High but that doesn't seem to be possible with validation attributes.

[Cmdlet(VerbsCommon.Get, "FakeData")]
public class GetFakeData : PSCmdlet
{
    [Parameter(Mandatory = true)]
    [ValidateNotNullOrEmpty]
    public int Low { get; set; }

    [Parameter(Mandatory = true)]
    [ValidateNotNullOrEmpty]
    public int High { get; set; }

    protected override void BeginProcessing()
    {
        if (Low >= High)
        {
            // Is there a better exception to throw here?
            throw new CmdletInvocationException("Low must be less than High");
        }

        base.BeginProcessing();
    }

    protected override void OnProcessRecord()
    {
       // Do stuff...
    }
}

Is there is a better way to do this? The main thing I don't like about the solution above is that I can't throw a ParameterBindingException like the validation attributes would do since it's an internal class. I could throw ArgumentException or PSArgumentException but those are really for Methods not cmdlets.

Answer 1

You need something like in the cmdlet get-random. Because you can't use [validatescript()] attribute in a cmdlet 'cause it's valid only for powershell function/script at run-time you need to steal the idea from microsoft.powershell.utility\get-random:

The value check is done in the BeginProcessing() and use a customized error ThrowMinGreaterThanOrEqualMax

protected override void BeginProcessing()
    {
      using (GetRandomCommand.tracer.TraceMethod())
      {
        if (this.SetSeed.HasValue)
          this.Generator = new Random(this.SetSeed.Value);
        if (this.EffectiveParameterSet == GetRandomCommand.MyParameterSet.RandomNumber)
        {
          if (this.IsInt(this.Maximum) && this.IsInt(this.Minimum))
          {
            int minValue = this.ConvertToInt(this.Minimum, 0);
            int maxValue = this.ConvertToInt(this.Maximum, int.MaxValue);
            if (minValue >= maxValue)
              this.ThrowMinGreaterThanOrEqualMax((object) minValue, (object) maxValue);
            this.WriteObject((object) this.Generator.Next(minValue, maxValue));
          }
          else
          {
            double min = this.ConvertToDouble(this.Minimum, 0.0);
            double max = this.ConvertToDouble(this.Maximum, double.MaxValue);
            if (min >= max)
              this.ThrowMinGreaterThanOrEqualMax((object) min, (object) max);
            this.WriteObject((object) this.GetRandomDouble(min, max));
          }
        }
        else
        {
          if (this.EffectiveParameterSet != GetRandomCommand.MyParameterSet.RandomListItem)
            return;
          this.chosenListItems = new List<object>();
          this.numberOfProcessedListItems = 0;
          if (this.Count != 0)
            return;
          this.Count = 1;
        }
      }
    }

...

private void ThrowMinGreaterThanOrEqualMax(object min, object max)
    {
      if (min == null)
        throw GetRandomCommand.tracer.NewArgumentNullException("min");
      if (max == null)
        throw GetRandomCommand.tracer.NewArgumentNullException("max");
      string errorId = "MinGreaterThanOrEqualMax";
      this.ThrowTerminatingError(new ErrorRecord((Exception) new ArgumentException(this.GetErrorDetails(errorId, min, max).Message), errorId, ErrorCategory.InvalidArgument, (object) null));
    }

You can use a decompiler ( dotPeak ) to see the rest of the code to learn more on custom error for cmdlet