Unicode support for Invoke-Sqlcmd in PowerShell

Question

The PowerShell sqlps module provides core support for SQL Server access from within PowerShell and its Invoke-Sqlcmd cmdlet is its main workhorse for executing literal queries or SQL script files (analogous to the non-PowerShell sqlcmd utility). I recently tried some experiments to confirm that Invoke-Sqlcmd handles Unicode and had some surprising results.

I started with this simple script file (named unicode.sql):

CREATE TABLE #customers

( [IdCust] int,
  [FirstName] nvarchar(25),
  [SurName] nvarchar(25)
);
INSERT INTO #customers VALUES (4, N'Hans', N'Grüßner')
SELECT * FROM #customers;
DROP TABLE #customers;

Note that the surname has some typical Unicode characters one might find in a German name, for example.

---

Results

SQL Server Management Studio: Renders correctly when output to grid or to text, e.g.

IdCust      FirstName                 Surname
----------- ------------------------- -------------------------
4           Hans                      Grüßner

sqlcmd utility: Renders correctly whether run from a DOS shell or a PowerShell, e.g.

C:\> sqlcmd -S .\SQLEXPRESS -i unicode.sql

IdCust      FirstName                 Surname
----------- ------------------------- -------------------------
          4 Hans                      Grüßner

PowerShell Invoke-Sqlcmd: Renders incorrectly (whether output as text as shown below or piped into Out-Gridview):

PS> Invoke-Sqlcmd -Server .\sqlexpress -InputFile unicode.sql

IdCust FirstName           Surname
------ ---------           -------
     4 Hans                Gr??ner

The MSDN documentation for Invoke-Sqlcmd mentions Unicode only in passing, comparing its command-line switches with those of sqlcmd, showing that while the latter has a -u option for outputting Unicode (which was not even needed in my experiment above), Invoke-Sqlcmd has no equivalent parameter.

I have found nothing at all regarding this point through extensive web searching but I still hold out hope that this is in some way a user error on my part. Is there a way to preserve the input data when retrieving it with Invoke-Sqlcmd in PowerShell?

Answer 1

Update I tested invoke-sqlcmd on another machine and it works, so maybe the rest of this doesn't apply...

Update 2 Only seems to have issue with -inputfile when executing via -Query parameter invoke-sqlcmd works fine.

From what I can tell this has something to do with ADO.NET DataTable when converting a string. It works fine when you use an ExecuteScaler or ExecuteReader. Of course this doesn't fix invoke-sqlcmd, but does explain why:

$server = "$env:computername\sql1"
$database = "tempdb"
$query = @"
CREATE TABLE #customers

(     [SurName] nvarchar(25)
);
INSERT INTO #customers VALUES (N'Grüßner')
SELECT * FROM #customers;
"@


$connection=new-object System.Data.SqlClient.SQLConnection
$connection.ConnectionString="Server={0};Database={1};Integrated Security=True" -f $server,$database
$command=new-object system.Data.SqlClient.SqlCommand($query,$connection)
$connection.Open()
$command.ExecuteScalar()
$connection.Close()

Update 3 The encoding of the file seems to be the key. Looking at [System.IO.File]::ReadAllText, the MSDN doc states it will only detect UTF-8 or UTF-32 encoding. http://msdn.microsoft.com/en-us/library/ms143369(v=vs.90).aspx

If I save the .sql file with UTF-8, using the -inputfile param works. You can choose UTF-8 when saving .sql file in SSMS, but here's some Powershell code to check and change the encoding also. You'll need to grab Get-FileEncoding.ps1 from http://poshcode.org/2075

. .\Get-FileEncoding.ps1 
Get-FileEncoding -Path E:\bin\unicode.sql

$query = get-content E:\bin\unicode.sql
$query= $query -join "`n"
$query | Out-File -FilePath e:\bin\unicode.sql -Encoding UTF8 -force

Get-FileEncoding -Path E:\bin\unicode.sql

Answer 2

When executing SQLCMD, you must specify the encoding.

EXEC xp_cmdshell 'for %f in ("{Dir}*.sql") do sqlcmd -S {Server} -U {username} -P {password} -d {database} -i "%f" -b -f 65001'