PowerShell Export-Csv No Headers

Question

I'm attempting to create multiple CSV files with no headers using Export-Csv within a PowerShell script. I've read a few posts about using the Select-Object -Skip 1 trick, but I haven't gotten it to successfully work on my code. Here's how I tried to call it:

$DataSet.Tables[0] | Select-Object -Skip 1 |
    Export-Csv -NoTypeInformation -Path "C:\$($csvname[$i])_$timer.csv"

Here's my full, functioning code:

function Run-Query {
    Param([string[]]$queries,[string[]]$csvname)
    Begin {
        $SQLServer = 'myserver'
        $Database = 'mydatabase'
        $SqlConnection = New-Object System.Data.SqlClient.SqlConnection
        $SqlConnection.ConnectionString = "Server = $SQLServer; Database = $Database; Integrated Security = True"
        $timer = (Get-Date -f yyyy-MM-dd)
    } # End Begin
    Process {
        # Loop through each query
        for ($i = 0; $i -lt $queries.Count; $i++) {
            $SqlCmd = New-Object System.Data.SqlClient.SqlCommand

            # Use the current index ($i) to get the query
            $SqlCmd.CommandText = $queries[$i]
            $SqlCmd.Connection = $SqlConnection
            $SqlAdapter = New-Object System.Data.SqlClient.SqlDataAdapter
            $SqlAdapter.SelectCommand = $SqlCmd
            $DataSet = New-Object System.Data.DataSet
            $SqlAdapter.Fill($DataSet)

            # Use the current index ($i) to get the sheetname for the CSV
            $DataSet.Tables[0] | Export-Csv -NoTypeInformation -Path "C:\$($csvname[$i])_$timer.csv"
        }
    } # End Process
    End {
        $SqlConnection.Close()
    }
} # End function run-query.

#Entery Query
$queries = @()
$queries += @'
SELECT * FROM table2
'@
$queries += @'
SELECT * FROM table1
'@

#List of CSV file names.
$csvname = @()
$csvname += 'file1'
$csvname += 'file2'
Run-Query -queries $queries -csvname $csvname

Answer 1

You're misunderstanding how CSVs work in PowerShell. Assume you have a CSV file with the following content:

foo,bar,baz
A,1,2
B,3,4

Importing that CSV via Import-Csv will give you 2 objects with the properties foo, bar, and baz filled with the values from one data row each. Using JSON notation the list of objects would look like this:

[
    {
        "foo": "A",
        "bar": "1",
        "baz": "2"
    },
    {
        "foo": "B",
        "bar": "3",
        "baz": "4"
    }
]

Export-Csv does basically the same, just in reverse. I takes objects as input and writes the values of their properties to the fields of the CSV. The fields are determined by the properties of the first object Export-Csv receives. If one of these properties is missing in a subsequent object a NULL value is written to the respective field. If a subsequent object has additional properties that the first object didn't have they are ignored.

Using Select-Object -Skip 1 with either Import-Csv or Export-Csv is useless, because normally you don't want to skip any of the input or output objects (otherwise you'd lose data).

There are, however, two other cmdlets similar to Import-Csv and Export-Csv that read and write strings rather than files: ConvertFrom-Csv and ConvertTo-Csv. Import-Csv is basically a combination of Get-Content and ConvertFrom-Csv, and Export-Csv a combination of ConvertTo-Csv and Set-Content.

So, where does Select-Object -Skip 1 come into play here? When you convert your object input to CSV via ConvertTo-Csv and skip the first row of the string output, you effectively remove the header line from the output text, something you couldn't do with Export-Csv.

Demonstration:

PS C:\> $data = @()
PS C:\> $data += [PSCustomObject]@{"foo"="A"; "bar"="1"; "baz"="2"}
PS C:\> $data += [PSCustomObject]@{"foo"="B"; "bar"="3"; "baz"="4"}
PS C:\> $data | Format-Table -Auto

foo bar baz
--- --- ---
A   1   2
B   3   4

PS C:\> $data | ConvertTo-Csv -NoType
"foo","bar","baz"
"A","1","2"
"B","3","4"
PS C:\> $data | ConvertTo-Csv -NoType | Select-Object -Skip 1
"A","1","2"
"B","3","4"

Pipe that output into Set-Content and you have the headerless CSV files you want.

$DataSet.Tables[0] |
    ConvertTo-Csv -NoType |
    Select-Object -Skip 1 |
    Set-Content "C:\$($csvname[$i])_$timer.csv"

Answer 2

Here's one way to output CSV input objects as tab-separated data without headers:

function ConvertTo-Tsv {
  param(
    [parameter(ValueFromPipeline=$true)]
    [Object[]] $Object
  )
  process {
    foreach ( $objectItem in $Object ) {
      ($objectItem.PSObject.Properties | Select-Object -ExpandProperty Value) -join "`t"
    }
  }
}

You could then do either of the following:

Import-Csv "inputfile.csv" | ConvertTo-Tsv

or

$data = Import-Csv "inputfile.csv"
ConvertTo-Tsv $data