Having trouble getting connections

I have a custom Data Flow Destination Adapter that is looking for a particular type of ConnectionManager. I want to check that this connection exists in the Validate method. I first checked the ComponentMetedata.RuntimeConnections, but it was was an empty collection. I am guessing that it gets populated at runtime. Is there anything available for to check at design time?

Another way of asking this would be, Is there a way to programatically select which ConnectionManager gets assigned to a RuntimeConnection? Normally, this is done on the Connections tab of the Advanced Editor.

Thanks,

Graham

You can assign a connection manager at design time to your adapter and check that it is the correct type in the Validate() method. You will need to let your component know that it is expecting a connection manager.

Allan

|||I realize that I can do that by opening the advanced editor on my destination task and selecting the connection; but I dont want to do it that way. I want to Programmatically look for a connection of a specific type. Below is a code snippet of what I tried to do, but that doesnt seem to work. The connection is not set.

//this takes place in the onload handler for my DestinationAdapter form
//connections is passed in from the UI class that implements IDtsComponentUI
ConnectionManager connMan = null;
foreach (ConnectionManager cm in connections)
{
if (cm.InnerObject.GetType() == typeof(ProfileConnectionManager))
{
connMan = cm;
}
}
profConnMan = ProfileConnectionManager.FindProfileConnection(connections);
dtsComponentMetaData.RuntimeConnectionCollection[0].ConnectionManager = DtsConvert.ToConnectionManager90(connMan);

|||Is there a reason you don't just get the value of ConnectionManagerType on the properties collection? Also, you're attempting to use a managed code idiom for a connection that may not be managed. Most of the connection managers are native. Is this a custom connection manager? I don't know about the ProfileConnectionManager.|||

You may also want to take a look at IDtsConnectionService.GetConnections and IDtsConnectionService.CreateConnection. Useful for writing UIs. They all work around the connection type that Kirk mentions.

|||ProfileConnectionManager is a custom connection that I have written. I want to allow a custom data flow destination adapter that I have written to look for a ProfileConnectionManager and if one is defined, add it to my CustomDestinationAdapter.ComponentMetaData.RuntimeConnections collection.

In normal situations, this is done at design time by openning the advanced editor and selecting the connection manager from a dropdown. In my situation, I am allowing only one ProfileConnectionManger to be created. So if it exists, I dont want the user to have to select it from the advanced editor; I want that to be done programmatically by me.

When you refer to properties collection, are you saying I should add a property to MyCustomDestinationAdapter.ComponentMetaData.CustomPropertyCollection and assign my custom connection as its value?

thanks,
Graham|||

I would use GetConnections, albeit in a UI, something like this-

foreach (ConnectionManager connectionManager in _dtsConnectionService.GetConnectionsOfType(connectionType))
{
comboBox.Items.Add(connectionManager.Name);
}

connectionType is the string that identifies the type of connection I want. For example if I want ADO.Net SQLClient connections I need to use-

"ADO.NET:System.Data.SqlClient.SqlConnection, System.Data, Version=2.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089";

It is a bit annoying that I have to use the fully qualified name, but that's what works.

The ConnectionManager also has a CreationName property which I use often for validation, and Kirk suggests a ConnectionType property, so using this you could loop as you where above, and check such a property to see if it is your connection. What you should not do is try and get a .Net Type object for the connection manager, as not all of them are .Net objects.

sql

Having problems using Breakpoint in Script Task.

I am still pretty new to SSIS, SQL, and DOT NET. I came from the UNIX world. I have a SSIS package. On the Control Flow one of the Items I have is a Script Task. I was able to successfully set breakpoints in the Script Task and they worked fine.I could step through the script and check values in variables.Life was good.

Now something happened.I set the breakpoint, from the menu I select “start with debugging” and I get the following window:

-- -

Visual Studio Just-In-Time Debugger

An unhandled exception (‘System.Runtime.InteropServices.COMException’) occurred in DTAttach.exe [3380].

Possible Debuggers:

New instance of Microsoft CLR Debugger 2003

New instance of Visual Studio .NET 2003

New instance of Visual Studio 2005

[_] set the currently selected debugger as the default

[_] Manually choose the debugging engines

Do you want to debug using the selected debugger?

--

I have tried selecting Yes, but that doesn’t work.If I delete all breakpoints the package runs fine.I greatly appreciate your help.

You are not alone.

http://forums.microsoft.com/MSDN/ShowPost.aspx?PostID=1846879&SiteID=1|||Thanks Phil. I found a work around for my problems and posted it on the thread you provided the link to above.

Handling flat files that do not exist

Hello,

I have a package that contains 22 data flow tasks, one for each flat file that I need to process and import. I decided against making each import a seperate package because I am loading the package in an external application and calling it from there.

Now, everything works beautifully when all my text files are exported from a datasource beyond my control. I have an application that processes a series of files encoded using EBCDIC and I am not always gauranteed that all the flat files will be exported. (There may have not been any data for the day.)

I am looking for suggestions on how to handle files that do not exist. I have tried making a package level error handler (Script task) that checks the error code ("System::ErrorCode") and if it tells me that the file cannot be found, I return Dts.TaskResult = Dts.Results.Sucsess, but that is not working for me, the package still fails. I have also thought about progmatically disabling the tasks that do not have a corresponding flat file, but it seems like over kill.

So I guess my question is this; if the file does not exist, how can I either a) skip the task in the package, or b) quietly handle the error and move on without failing the package?

Thanks!

Lee.

I recall that I have a couple data flows in my package, each within a sequence container and connected them with completion(blue) arrows and even if the first one failed the second one continued on...

|||

That is one solution that I could use. It doesn't seem like it is the "elegant" way to handle it, but it will work.

Thanks for the suggestion.

|||

well

Is dragging red arrow to some alert/send mail task will not serve purpose?

|||

Scoutn wrote:

Hello,

I have a package that contains 22 data flow tasks, one for each flat file that I need to process and import. I decided against making each import a seperate package because I am loading the package in an external application and calling it from there.

Now, everything works beautifully when all my text files are exported from a datasource beyond my control. I have an application that processes a series of files encoded using EBCDIC and I am not always gauranteed that all the flat files will be exported. (There may have not been any data for the day.)

I am looking for suggestions on how to handle files that do not exist. I have tried making a package level error handler (Script task) that checks the error code ("System::ErrorCode") and if it tells me that the file cannot be found, I return Dts.TaskResult = Dts.Results.Sucsess, but that is not working for me, the package still fails. I have also thought about progmatically disabling the tasks that do not have a corresponding flat file, but it seems like over kill.

So I guess my question is this; if the file does not exist, how can I either a) skip the task in the package, or b) quietly handle the error and move on without failing the package?

Thanks!

Lee.

You can use a script task to check whether or not a file exists. This code should help:

File.Exists Method
(http://msdn2.microsoft.com/en-us/library/system.io.file.exists.aspx)

-Jamie

|||

Jamie,

First allow me to say that your blog has been an invaluable resource. I have learned a lot from your articles, thank you!

Utsav, I am not looking to be notified of the error; I just want the task to merrily carry on like nothing took place. I do understand what you meant though, thank you.

Sometimes the flat file will be there and sometimes it won't be. If it isn't, I would still like the task move on to the next with a "Success" result, not just completion. That is why I was thinking of disabling the tasks that do not have corresponding flat files progmatically when I load the package from my application.

I guess I should be asking; when I set Dts.TaskResult = Dts.Results.Success, why does it not actually return success?

(I know there are a couple of ways I can get around this issue, I'm just looking for the easiest without setting the task to continue on "Completion" ;)

Thanks again,
Lee.

|||

Check if this thread has something that can help you.

http://forums.microsoft.com/MSDN/ShowPost.aspx?PostID=859625&SiteID=1

Instead of disabling the task; try to use an expression in the precedence constraint in the control flow.

Rafael Salas

|||Thank you very much. I did search the forum, I guess I was just using the wrong search term.

Perfect!|||

The thread (and the subsequent blog articles) solved my problem. I needed to put each task in a sequence container with their own script task checking for the files; if I used one script task to look for all the files the next task would not run because at least one file would return false.

Thanks again for everyones help!