When I run Dave Sexton's code, I get:
Select-Object : Cannot convert System.Management.Automation.PSObject to one of the following
types {System.String, System.Management.Automation.ScriptBlock}.
At line:15 char:12
+ $x | Select <<<< $f | Export-Csv ColsRem.test.$time.csv -NoTypeInformation
+ CategoryInfo : InvalidArgument: (:) [Select-Object], NotSupportedException
+ FullyQualifiedErrorId :
DictionaryKeyUnknownType,Microsoft.PowerShell.Commands.SelectObjectCommand
I corrected this issue by adding one more line, to force each array element to be a string.
$x = Import-Csv YourFile.csv
$f = $x[0] | Get-Member -MemberType NoteProperty | Select name
$f | Add-Member -Name count -Type NoteProperty -Value 0
$f | %{
$n = $_.Name
$_.Count = @($x | Select $n -ExpandProperty $n | ? {$_ -ne ''}).count
}
$f = @($f | ? {$_.count -gt 0} | Select Name -expandproperty Name)
# I could get the select to work with strings separated by commas, but the array would
# always produce the error until I added the following line, explicitly changing the
#values to strings.
$f = $f | Foreach-Object { "$_" }
$x | Select $f | Export-Csv NewFile.csv -NoTypeInformation
My import CSV contains a few hundred columns and about half likely won't be populated, so getting rid of the extra columns was necessary. Now I just need to figure out how to counteract the unintended re-ordering of the columns into alphabetical order by name, without changing the names.