I'm calling a powershell script like so, from within another script.
function test_execute_sql_2_direct_query()
{
& C:\Server\Scripts\PowerShellDeploy\Execute-SQL-Command.ps1 -SERVER_NAME "myserver" -SQL_COMMAND_TYPE "input" -SQL_FILE ".\select.sql" -USER_NAME "username" -PASSWORD "password" -DATABASE "accounts"
}
When I call that function I get an error message from Powershell that says 'Cannot validate argument on parameter 'Query'. The argument is null or empty....'
I don't quite follow why I'm getting that message? The Execute-SQL-Command script doesn't have an argument called 'Query'. Quite possible I'm missing something.
Here's its paramenter section:
param(
[Parameter(
HelpMessage="Server Name")]
[string]$SERVER_NAME = "",
[Parameter(
HelpMessage="Database")]
[string]$DATABASE = "",
[Parameter(
HelpMessage="Path to .SQL File to run. Used in input file mode (not direct query mode)")]
[string]$SQL_FILE ="",
[Parameter(
HelpMessage="Server Port (default = 1433)")]
[string]$SERVER_PORT = "1433",
[Parameter(
HelpMessage="Server Transport (default = tcp")]
[string]$TRANSPORT = "tcp",
[Parameter(
HelpMessage="Method of connecting to database <passthrough, credentialed>")]
[string]$CONNECT_METHOD = "passthrough",
[Parameter(
HelpMessage="Username -- used only for credentialed login")]
[string]$USER_NAME = "username",
[Parameter(
HelpMessage="Command type-- direct query or input file (.sql)")]
[string]$SQL_COMMAND_TYPE = "DIRECT_QUERY",
[Parameter(
HelpMessage="SQL statement text type-- used in direct queries")]
[string]$SQL_STATEMENT = "",
[Parameter(
HelpMessage="Password -- used only for credentialed login")]
[string]$PASSWORD = "pwd")
This one is chalked up to user error-- the actual error was coming from within a line called in Execute-SQL-Command.ps1-- specifically, the Invoke-Sqlcmd cmdlet.
Related
I have the following code:
Function test {
Param (
[Parameter(Mandatory=$True, ValueFromPipeline=$True, ParameterSetName="p1", position=0)]
[string]$sFile,
[Parameter(Mandatory=$True, ValueFromPipeline=$True, ParameterSetName="p2", position=0)]
[int]$iFile,
[string]$Secret = "s3cr3t"
)
BEGIN {
"[BEGIN]Secret=${Secret}"
}
PROCESS {
"[PROCESS]Secret=${Secret}"
}
}
If i execute:
"file.txt" | test
I get:
[BEGIN]Secret=
[PROCESS]Secret=s3cr3t
My question is, why secret is empty in the begin block?
I am trying to perform some initialization actions in the begin block based on Secret and this behavior is stopping me.
Note that when i run test without pipeline input, the secret is initialized at BEGIN bock as well. Finally my powershell version is 2.
I know how to add switch parameters to PowerShell scripts, but the name of the switch is always a (valid) identifier like -Enable because the name is generated from the backing PowerShell variable.
[CmdletBinding()]
param(
[switch] $Enable = $false
)
Some tools have switches like -2008. Normally, one would name the switch $2008 but this is not a valid identifier.
How can I implement such a switch as a boolean value in PowerShell?
Or in other words: How to specify a different parameter name then the backing variable?
Edit 1
I wasn't aware of number only variables (which is very strange for a programming language...). Anyhow, I created that example:
function foo
{ [CmdletBinding()]
param(
[switch] $2008
)
Write-Host "2008=$2008"
}
For this code, which is accepted as valid PowerShell, I get a auto completion as wanted. But when providing that parameter, I get this error message:
foo : Es wurde kein Positionsparameter gefunden, der das Argument "-2008" akzeptiert.
In Zeile:1 Zeichen:1
+ foo -2008
+ ~~~~~~~~~
+ CategoryInfo : InvalidArgument: (:) [foo], ParameterBindingException
+ FullyQualifiedErrorId : PositionalParameterNotFound,foo
Translation: No positional parameter was found, that accepts the argument "-2008".
The purpose of the script is to provide translations / wrappers for command line interfaces.
Here is a set of standardized executables with parameters:
vcom.exe -2008 design.vhdl
vsim.exe -2008 design
Translation:
ghdl.exe -a --std=2008 design.vhdl
ghdl.exe -e --std=2008 design
ghdl.exe -r --std=2008 design
I would like to keep the feature of auto completion for parameters, otherwise I could process all remaining parameters and translate them by hand.
PowerShell doesn't support numeric parameters (See this answer).
Could a validateset be an acceptable solution to your problem ?
Validate set does benefit from the autocompletion feature and this is the next best thing, in my opinion, of what you wanted.
Foo & Foos — both are the same, except Foos accept multiple parameters.
foo -std 2008
foos -std 2008,2009,2010
function foo
{ [CmdletBinding()]
param(
[ValidateSet("2008",
"2009",
"2010",
"2011","2012")][INT]$std
)
Write-Host "std:$std"
}
function foos
{ [CmdletBinding()]
param(
[ValidateSet("2008",
"2009",
"2010",
"2011","2012")][INT[]]$std
)
$std | foreach {Write-Host "std: $_"}
}
foo -std 2008
foos -std 2008,2009,2010
I've got a release definition in vsts that needs access to a password I've defined as a secret variable in a variable group. I've linked that group to this release definition but when I run the task the parameter ends up blank. Is there anything special I need to do to get the value of a secret variable?
Definition of the powershell task that uses the password
Linked variable group
Error output:
2017-08-09T14:01:17.9262375Z ##[command]. 'C:\agent_work\r1\a\$(AGENT.BUILDDIRECTORY)\RCV\deploys\common\Get-FromArtifactory.ps1' -repoUsername developer -repoPassword -repoPath libs-snapshot-local
2017-08-09T14:01:18.8168538Z ##[error]C:\agent_work\r1\a\$(AGENT.BUILDDIRECTORY)\RCV\deploys\common\Get-FromArtifactory.ps1 : Missing an argument for parameter 'repoPassword'. Specify a parameter of type 'System.String' and try again.
Edited to add more info as requested.
This is the start of my Get-FromArtifactory.ps1
[CmdletBinding()]
param (
[Parameter(Mandatory)]
[String]
$repoUsername,
[Parameter()]
[String]
$repoPassword,
# Other params
)
#setup credentials object for arty access
$secPassword = $repoPassword | ConvertTo-SecureString -asPlainText -Force
$credential = New-Object System.Management.Automation.PSCredential($repoUsername, $secPassword)
Edit 15/08/2017 - I've updated it to use quotes as suggested by #Marina-MSFT but it's still just passing in blank.
Amended script to print out part of password:
[CmdletBinding()]
param (
[Parameter(Mandatory)]
[String]
$repoUsername,
[Parameter()]
[String]
$repoPassword,
#other params
)
write-host "pword $repoPassword"
write-host "1st char: $($repoPassword[0])"
2017-08-15T09:23:00.8606081Z ##[command]. 'C:\agent_work\r1\a\RCV\deploys\common\Get-FromArtifactory.ps1' -repoUsername developer -repoPassword "" -repoPath libs-snapshot-local -artifactName ipo-deposit-accounts_2.11 -artifactPath uk/gov/ipo -artifactVersion 0.1-SNAPSHOT -downloadDir C:\agent_work\r1\a
2017-08-15T09:23:02.3606174Z pword
2017-08-15T09:23:02.3606174Z 1st char:
Same with single quotes
2017-08-15T09:22:10.6573655Z ##[command]. 'C:\agent_work\r1\a\RCV\deploys\common\Get-FromArtifactory.ps1' -repoUsername developer -repoPassword '' -repoPath libs-snapshot-local -artifactName ipo-deposit-accounts_2.11 -artifactPath uk/gov/ipo -artifactVersion 0.1-SNAPSHOT -downloadDir C:\agent_work\r1\a
2017-08-15T09:22:11.2198723Z pword
2017-08-15T09:22:11.3761178Z 1st char:
I've got it working now, there were 2 things going on.
Having . in the name of a secret variable in a variable group
doesn't work. I recreated this issue in a separate release
definition/variable group.
Even after I removed the . in my var
names I still had to recreate the variable group for it to work.
Bug for issue with . in variable name raised here: https://developercommunity.visualstudio.com/content/problem/94475/secret-variable-in-variable-not-group-not-availabl.html
I am writing a PowerShell function which can take pipeline input. Specifically I am testing it with Import-CSV. Many of the params are not mandatory, which means sometimes the CSV will not have those columns. For boolean values this is working as expected, but with string values, a missing CSV field yields a copy of the row object in the string field.
Here is an example problem parameter:
[Parameter(Mandatory=$False,
ValueFromPipeline=$True,
ValueFromPipelinebyPropertyName=$True,
HelpMessage="TCP Port of the remote server")]
[Alias('Port', 'Remote Server Port')]
[string]$RemoteServerPort = "5500",
Now, if the field is missing, I would expect the value to be "5500" as specified, but instead I get:
$RemoteServerPort = #{Name=KG; IP=10.1.1.1; Username=Admin; Password=}
I've done some looking around, but frankly I'm not even sure what to search for.
This is because you specified ValueFromPipeline=$True so that PoSh coerces the piped object to a string if it cannot bind the parameter by property name. You could solve that by removing ValueFromPipeline=$True from this parameter and introduce another one to be bound to the piped object, i.e. something like this
function MyTestFunc() {
param(
[Parameter(ValueFromPipeline=$True)]
[object]$PipedObj,
[Parameter(Mandatory=$False,
ValueFromPipelinebyPropertyName=$True,
HelpMessage="TCP Port of the remote server")]
[Alias('Port', 'Remote Server Port')]
[string]$RemoteServerPort = "5500"
)
Write-Host "Port: $RemoteServerPort / Obj: $PipedObj"
}
$o1 = [pscustomobject]#{"Server" = "123"; "Port" = "12345"}
$o2 = [pscustomobject]#{"Server" = "1234"; "OS" = "Win3.1"}
$o1 | MyTestFunc
$o2 | MyTestFunc
Will result in
Port: 12345 / Obj: #{Server=123; Port=12345}
Port: 5500 / Obj: #{Server=1234; OS=Win3.1}
A way to see in detail what is actually happening behind the scenes is to use Trace-Command like so
Trace-Command ParameterBinding {$o2 | MyTestFunc} -PSHost
I have a powershell module which attempts to upload a blob to azure storage. Everything checks out until the last line which actually uploads the blob.
I receive the following error:
Exception calling "UploadText" with "1" argument(s):
"The specified resource does not exist."
At line:1 char:1
+ $blob.UploadText("asdasdfsdf")
+ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+ CategoryInfo : NotSpecified: (:) [], MethodInvocationException
+ FullyQualifiedErrorId : StorageClientException
I have also tried using the overload with 3 args, but the same issue exists there as well.
Here is the module:
Function Add-BlobText
{
[CmdletBinding()]
param(
[Parameter(Mandatory = $true,Position = 0)]
[string]
$StorageAccount,
[Parameter(Mandatory = $true,Position = 1)]
[string]
$Container,
[Parameter(Mandatory = $true,Position = 2)]
[string]
$BlobName,
[Parameter(Mandatory = $true, ValueFromPipeline = $true)]
[string]
$BlobText
) #end param
Add-Type -Path "C:\Assemblies\Microsoft.WindowsAzure.StorageClient.dll"
Set-AzureSubscription -SubscriptionName "MySubName"
$secondaryKey = (Get-AzureStorageKey -StorageAccountName $storageAccount).Secondary
$creds = New-Object Microsoft.WindowsAzure.StorageCredentialsAccountAndKey($StorageAccount,$secondaryKey)
$cloudStorageAccount = New-Object Microsoft.WindowsAzure.CloudStorageAccount($creds, $true)
[Microsoft.WindowsAzure.StorageClient.CloudBlobClient]$cloudBlobClient = New-Object Microsoft.WindowsAzure.StorageClient.CloudBlobClient($cloudStorageAccount.BlobEndpoint)
[Microsoft.WindowsAzure.StorageClient.CloudBlobContainer]$blobContainer = $cloudBlobClient.GetContainerReference($Container)
[Microsoft.WindowsAzure.StorageClient.CloudBlob]$blob = $blobContainer.GetBlobReference($BlobName)
$blob.UploadText($BlobText)
} #end Function Add-BlobText
Update:
I have been able to get this working as a binary module (below). If anyone can figure out why UploadText() works within a binary module but throws an exception in a script module, please let me know.
[Cmdlet(VerbsCommon.Add, "BlobText")]
public class AddBlobText : PSCmdlet
{
[Parameter(Mandatory = true, Position = 0)]
public string StorageAccount { get; set; }
[Parameter(Mandatory = true, Position = 1)]
public string Container { get; set; }
[Parameter(Mandatory = true, Position = 2)]
public string BlobName { get; set; }
[Parameter(Mandatory = true, ValueFromPipeline = true)]
public string BlobText { get; set; }
protected override void ProcessRecord()
{
PowerShell ps = PowerShell.Create();
ps.AddScript("Set-AzureSubscription -SubscriptionName 'MySubName'");
string keyScript = "( Get-AzureStorageKey -StorageAccountName " + StorageAccount + " ).Secondary";
ps.AddScript(keyScript);
Collection<PSObject> result = ps.Invoke();
string secondaryKey = result[0].ToString();
StorageCredentialsAccountAndKey credentials = new StorageCredentialsAccountAndKey(StorageAccount, secondaryKey);
CloudStorageAccount storageAccount = new CloudStorageAccount(credentials, true);
CloudBlobClient blobClient = storageAccount.CreateCloudBlobClient();
CloudBlobContainer container = blobClient.GetContainerReference(Container);
var blob = container.GetBlobReference(BlobName);
blob.UploadText(BlobText);
}
}
This is probably because your container does not exist. You should call CreateIfNotExist after initializing the container to make sure it exists:
[Microsoft.WindowsAzure.StorageClient.CloudBlobContainer]$blobContainer = $cloudBlobClient.GetContainerReference($Container)
$blobContainer.CreateIfNotExist() <-- Here
[Microsoft.WindowsAzure.StorageClient.CloudBlob]$blob = $blobContainer.GetBlobReference($BlobName)
$blob.UploadText($BlobText)
This error is very ambiguous and misleading but there are instances' where Azure Storage can get "confused". Looking at Sandrino's example and specifically this line,
[Microsoft.WindowsAzure.StorageClient.CloudBlob]$blob = $blobContainer.GetBlobReference($BlobName)
Not that Sandrino's answer is your issue but the exception you encountered will happen when passing a Url or possibly other confusing key strings to Azure Storage Containers.
Unfortunately I am not a Powershell guy but here is a reproducing example then fix in C#.
public void Save(string objId, T obj)
{
CloudBlob blob = this.container.GetBlobReference(objId); // Problematic if a URL
blob.Properties.ContentType = "application/json";
var serialized = string.Empty;
serialized = serializer.Serialize(obj);
if (this.jsonpSupport)
{
serialized = this.container.Name + "Callback(" + serialized + ")";
}
blob.UploadText(serialized);
}
Assume that this.container is a valid blob storage instance pointing to http://127.0.0.1:10000/devstoreaccount1/sggames or whatever you have for a valid container.
And objId is a key that contains a Url like https://www.google.com/accounts/o8/id?id=AItOawk4Dw9sLxSc-zmdWQHdZNcyzkTcvKUkhiE ...and yes this can happen, in my case this is an actual identity claim from Google using Azure ACS.
After the GetBlobReference call the blob instance has become corrupt which now looks at a messed up Uri -> https://www.google.com/accounts/o8/id?id=AItOawk4Dw9sLxSc-zmdWQHdZNcyzkTcvKUkhiE
Unfortunately the solution to simply call $blobContainer.CreateIfNotExist() is not a solution and wouldn't work. Key's that contain a Uri structure will simply be used to re-interpret the blob storage location.
The work around (other than daredev's Update) would be something like this:
if (Uri.IsWellFormedUriString(claim, UriKind.Absolute) && HttpUtility.ParseQueryString(claim).Count > 0)
{
claim = HttpUtility.ParseQueryString(claim)[0];
}
Add this code within my method above to clean up any Uri's, but you could use any appropriate method like Base64 encoding URLs if you need to maintain the full key.
Here are the before and after images showing the results as I described.
The bad:
notice the bad URI
this bad URI munged up the actual storage blob location
here is the same exception daredev had
The good:
the new scrubbed key, notice it's just the value on the URL's query string
Azure Storage URI looks good now
Eureka!
Hope this helps.
This is the PowerShell script I use to upload a file to Azure Blob: Uploading to Azure Storage
$SubscriptionName = ""
$SubscriptionId = ""
$DestContainer = ""
$StorageAccountName = ""
Import-AzurePublishSettingsFile -PublishSettingsFile "<Location of the publishsettings-file>"
Set-AzureSubscription -SubscriptionId $SubscriptionId -CurrentStorageAccountName $StorageAccountName
Select-AzureSubscription -SubscriptionName $SubscriptionName
Set-AzureStorageBlobContent -File "<File you want to upload>" -Container $DestContainer