The Foreach Loop container is one of the most flexible and useful controls available to SQL Server Integration Services developers, but the fact that it's relatively poorly documented means that it may not get used as often as it deserves.
In this example I'll describe how I used it to build a workaround for a problem in SQL Server log shipping. Due to low disk space we needed to delete log files as soon as they were no longer needed without changing the log maintenance configuration, but this had to be both automated and failsafe. The Foreach Loop container made this possible. You can adapt this solution for any situation where operations need to be regularly performed on an unknown number of files with no fixed filenames.
I'm assuming that you're familiar with using the SQL BI Dev Studio tools and building basic packages. If this isn't the case, I recommend working your way through the Integration Services Tutorial in SQL Books Online to become familiar with the basic concepts. Jamie Thomson's blog is an invaluable resource when you're ready to go further with SSIS.
Connections and Variables
I defined two connections in a new SSIS package, a regular SQL Server data source pointing towards the SQL log shipping monitor/target server and a connection pointing towards TEMPDB on the target SQL server.
I also defined four package-scoped variables (as shown below). These will be explained in more detail as we go on.
Setting the stage
The first step was an Execute SQL Task that creates a temporary table for the list of filenames we'll be handling. I used the TEMPDB connection and added this code to the SQLStatement property with a SQLSourceType of 'Direct input':
if not exists (select *
where xtype = 'U'
and name = 'tmpTLogFileList')
create table tmpTLogFileList
(id int identity(1, 1) not null,
FileName nvarchar(128) not null) else truncate table tmpTLogFileList
Building the list of files
The next step was to populate this table with a list of files found in a given folder. To do this I added a Foreach Loop Container to the package. On the Collection tab I selected the 'Foreach File enumerator' and entered the path to the folder. I could have used a variable here, but in this case the path was fixed so that wasn't necessary. The 'Name and extension' option was also selected.
On the 'Variable Mappings' tab I added a mapping to the varTLogFileName variable I defined earlier. The Foreach Loop will move through each file in the folder I defined and update this variable each time with each file name it encounters.
The container's properties were now correctly set. I then added a Execute SQL task within the container. This used the TEMPDB connection again with the following code:
insert into tmpTLogFileList (FileName)
On the 'Parameter Mapping' tab I added the varTLogFileName variable as an Input parameter with a mapping to the @FileName variable used in the SQL command (as shown above). At runtime, the Loop container runs this command once for each filename and the tmpTLogFileList SQL table is populated.
Checking which files can be safely deleted
I now had a table with a complete list of all the files found in the folder, but in this particular case I needed to be sure that these backup files had been correctly restored on the log shipping target server before deleting them.
I created a Data Flow task to build the list of files to be deleted. The source was an OLE DB connection to the log shipping target SQL server with a SQL query joining the tmpTLogFileList table with the system table used by SQL to store the restore status:
where id <=
(select id - 16
where FileName in
where succeeded = 1
and activity = 1))
and FileName like '%.TRN'
(The 'ID - 16' select merely ensures that the newest 16 backup files are always excluded from the delete in any case, as an additional safety measure).
The result set is a single-column list of filenames. I stored this in the Object variable varDeleteFileList that I'd created earlier using a Recordset Destination control with the 'VariableName' Custom Property set to varDeleteFileList as shown above.
Deleting the files
After this last step I now had a variable containing the list of files that could be safely deleted. To actually delete these files I added another Foreach Loop Container to my package.
This time I selected the 'Foreach ADO enumerator' as the enumerator type. I selected the varDeleteFileList variable from the ADO object source variable list and added a mapping to the varDeleteFileOName variable on the Variable Mappings tab. This is an Object-typed variable, allowing me to store a recordset. This recordset is then used to delete the files one-by-one as follows.
I added a Script task in the container. This was necessary in order to convert the value of the varDeleteFileOName variable from an Object to a string that could be used in the File System Task step (described below).
The Script Task was defined as follows:
The script code is short and simple:
Public Class ScriptMain
Public Sub Main()
Dim varStringFileName As String
varStringFileName = Dts.Variables("varDeleteFileOName").Value.ToString
Dts.Variables("varDeleteFileSName").Value = "\\SERVERNAME\TLog_Backups\" + varStringFileName
Dts.TaskResult = Dts.Results.Success
The path listed in the code is, of course, a hard-coded reference to the folder where the files to be deleted are stored but could also be populated by a variable if you so choose. The varDeleteFileSName variable is populated with a string-valued path and file name and the Script Task is complete.
The last step in the package was to add a File System Task inside the second Foreach Loop container. On the 'General' tab I selected 'Delete file' in the Operation drop-down, 'True' for IsSourcePathVariable and varDeleteFileSName as the SourceVariable.
And that's it! Running the package deletes all backup files that haven't been restored except the newest sixteen. I deployed the package to the production SSIS server and scheduled it with a regular SQL job to run hourly.
This is a fairly simple example that doesn't include much error-trapping, logging and so on, but I hope that it at least illustrates the power and flexibility of the Foreach Loop Container in SQL Server Integration Services.