Creating a Database Table
Creating a database and a table are common tasks that a DBA may undertake. You can use T-SQL with the Invoke-SqlCmd
cmdlet to do this, but a demonstration on how to do this with the SQL
Server PowerShell minishell using SMO is presented here to help you
better understand the new functionality that is available:
cd databases
$my_db=New-Object Microsoft.SqlServer.Management.Smo.Database
$my_db.Name="my_database"
$my_db.Parent=(Get-Item ..)
$my_db.Create()
cd my_database
cd tables
$my_tbl=New-Object Microsoft.SqlServer.Management.Smo.Table
$my_tbl.Name="my_table"
$my_tbl.Parent=(Get-Item ..)
$my_col=New-Object Microsoft.SqlServer.Management.Smo.Column
$my_col.Name="my_column"
$my_col.Parent=$my_tbl
$my_col.DataType= ([Microsoft.SqlServer.Management.Smo.DataType]::Int)
$my_tbl.Columns.Add($my_col)
$my_tbl.Create()
In
the preceding example, some new objects are created, some of their
properties are set, and some methods are called. You can search for the
particular SMO classes used in this example to gain further information.
In the future, there may be something like New-Database and New-Table cmdlets that help to create a database and table, which would likely reduce the preceding code to fewer than five lines.
Performing a Database Backup
Another example that may be useful is performing a database backup using SMO. Using the AdventureWorks2008R2 database, you can back up the database using just a few lines:
$server=New-Object Microsoft.SqlServer.Management.Smo.Server
$backup=new-object Microsoft.SqlServer.Management.Smo.Backup
$file=new-object Microsoft.SqlServer.Management.Smo.BackupDeviceItem
$file.Name="C:\backup\AW_DB.bak"
$backup.Devices.Add($file)
$backup.Database="AdventureWorks2008R2"
$backup.SqlBackup($server)
The preceding code could be copied into a .ps1 file (for this example, assume it’s copied to c:\temp\backup.ps1),
and it could be changed to accept two arguments. The preceding code
could be modified to the following code snippet so that it accepts
parameters from the command line:
param([string]$device=$(throw Write-Host "Device required"), [string]
$database=$(throw Write-Host "Database required"))
Write-Host "backup of $database to $device starting..."
$server=New-Object Microsoft.SqlServer.Management.Smo.Server
$backup=new-object Microsoft.SqlServer.Management.Smo.Backup
$file=new-object Microsoft.SqlServer.Management.Smo.BackupDeviceItem
$file.Name=$device
$backup.Devices.Add($file)
$backup.Database=$database
$backup.SqlBackup($server)
Write-Host "backup complete"
Get-Item $device
The changes in the preceding example introduce a new keyword, throw.
Without it, error messages would be thrown to the console, but this
might not help the end user to understand why it failed. For the purpose
of printing feedback to the console, the Write-Host and Get-Item cmdlets were also added to provide a limited amount of feedback and to finally provide the details of the final backup file.
Then to invoke the script, as
shown in the following example, you simply need to pass two parameters
to the script: the names of the file to back up to and the database to
actually back up.
PS>$backup_to="C:\backup\AdventureWorks2008R2.bak"
PS>$db="AdventureWorks2008R2"
PS> c:\temp\backup.ps1 $backup_to $db
As an example of using a
conditional statement in the preceding script, perhaps a check could be
done to see whether a backup has already been completed within the past
seven days. To accomplish this, you would add this particular section of
code just before the line Write-Host "backup of $database to $device starting...":
If((Test-Path $device) -and (Get-Item $device).LastWriteTime `
-gt (Get-Date).AddDays(-7)){
"Backup has been performed in last 7 days" Break
}
In the preceding code, a conditional statement is added to accomplish the check. The AddDays() method is passed a negative number which subtracts days from the current date..
Note
When you use the param
keyword, this section of code must be on the first noncommented line in
all scripts. Otherwise, the PowerShell parser returns an error.
Again, using SMO isn’t
necessarily for beginners, but the good thing is that scripts can easily
be created and passed around. The preceding example shows the bare
minimum required to do a database backup; several other options are
available, and the preceding code would actually overwrite the backup
each time it is run. There also isn’t any error checking of any sort,
which isn’t the best way to develop scripts.
Along with the New-Database and New-Table cmdlets that may come in the future, maybe a Start-DbBackup will be another good cmdlet to have available.
Checking Server Settings
From the SSMS, by right-clicking on the SQL Server
node and then starting a SQL Server PowerShell session, you can open a
console directly in the root of the default SQL Server in the example.
From here, you can easily obtain
information on the SQL Server. First, the location is set to the
instance that is to be queried, and then an object representing the SQL
Server is saved to a variable (this demonstrates the advanced features
mentioned earlier where objects can be saved to variables). The
properties of that variable can then be accessed as follows:
PS>Set-Location SQLSERVER:\SQL\<servername>\<instance_name>
PS>$sql_server=get-item .
PS>$sql_server.Information.VersionString
Using the Get-Member cmdlet discussed earlier, you can easily discover other members of the object contained in $sql_server, but this is left as an exercise for you to perform on your own.
Note
This example demonstrates an
important feature of the SQL Server provider: context sensitivity. In
the preceding example, the current location was in the root of the SQL
Server provider or database, and the command Get-Item.
was used. The dot in this command basically indicates that you want an
object that represents the current location in the provider. If the dot
were moved to a different location, this command would no longer work
the same way.
Checking the Database Usage
Using the object retrieved in the $sql_server variable, you can create a quick report of database usage using the databases property of that object, as shown here:
PS SQLSERVER:\SQL\<servername>\<instancename>>
$sql_server.databases| Format-Table -autosize Name,@{
Label= "% Used"
Expression={[math]::round((($_.spaceavailable/1kb)/$_.size),2)}
}Name % Used
---- ------
AdventureWorks2008R2 0.02
AdventureWorksDW2008R2 0
AdventureWorksLT2008R2 0.02
bigpubs2008 0.18
Customer 0.74
master 0.24
model 0.39
msdb 0.02
my_database 0.38
tempdb 0.8
Using the Format-Table
cmdlet, you can easily and quickly create all kinds of reports. Some
capabilities we haven’t discussed yet were used to create this report:
Calculated properties— The values displayed by Format-Table
can be calculated using scriptblocks. That allows the logic to be
highly customized. These scriptblocks are laid out as follows:
@{Label="some text value"
Expression={the scriptblock to evaluate here}
}
Direct access to the .NET Framework— The following line is directly from the .NET Framework:
"[math]::round(value,decimal)"
.NET functionality is being used to round out the numbers to the second decimal point.
PowerShell has special meaning for 1kb, 1mb, 1gb and 1tb, which all present the value of the counterpart in number of bytes—for example, 1kb=1024. The values can also be uppercase.
