I had an idea that it would be nice to be able to extend the functionality of an existing SSIS component or task by inheriting from it. Perhaps in a similar way to how it is possible to extend user controls in .Net.
e.g. The rowcount is a useful component but it would be good to create a new component that inherits from it and then override the PostExecute() method to fire an Information event containing the number of rows. That's a very simple example but I think you get the gist.
Does anyone think that would be useful? Or even made possible?
I was also wondering whether it would be possible for the SSI team to make all the icons used within SSIS available as .ico files so that we could modify them for tasks/components that might do similar things.
Whenever I make a breaking change to a custom SSIS component/tasks and update the Assembly Version, it seems to break my packages beyond repair, telling me it can't load the task:
Error loading Package1.dtsx: Error loading a task. The contact information for the task is "". This happens when loading a task fails.
All of the properties of said task now show:
Could not get value for property 'c-155-designer-name'. Specified cast is not valid.
Typically, a "breaking" change when it comes to code just means that you need to update your components to adhere to the new contract of the updated signatures. But with SSIS, it seems the only solution to this is to completely remove the component, and re-add the new version, and re-enter all of the property values/expressions. If I have a package containing 10 instances of a task that only had one property removed, for example, this results in a very time-consuming process of fixing my package.
So my questions:
1) Am I doing something wrong in my versioning/deployment that is causing my packages to unnecessarily break?
2) If this is just "by design" and the way it's meant to behave, what is the best practice for making breaking changes to custom tasks/components used by many packages? Should I just never change the assembly version, even when it is a breaking change (this seems to be less disastrous)?
3) As a last resort, if I'm stuck with having to fix the broken tasks, is there a better way to fix them rather than having to completely remove them, re-add them, and re-set all of their properties/expressions?
Everything I've read says that custom data flow components are built by inheriting from the Microsoft.SqlServer.Dts.Pipeline.PipelineComponent class.
But the stock components such as the Derived Column data flow transformation must each be implemented by their own class. So how do I base my custom components on those classes? The documentation for the PipelineComponent class doesn't list any such subclasses.
In several threads there has been discussion regarding adding connection managers to a package's data flow, etc. My challenge is that I have a large solution that contains many packages, and I need to change the connection manager linked to the data flow in all of the packages. When the solution was initially designed, data sources were used, and it has become a tedious maintenance issue to keep those in sync. We want to use a standard OLEDB connection manager, but adding a connection manager to each package and editing the corresponding data flow tasks in each package to use that new connection manager is a daunting task. I've coded a .Net module to access the packages, remove the old connection manager (data source) and add the new OLEDB data source. However, as I traverse the objects in the package hierarchy, when I come to the data flow object, the innerobject is not a dts object, but rather a _com object.. I can't seem to find any documentation/examples as to how to iterate the tasks within a data flow and change the connection manager. If you have any information, that would be quite helpful. If you reply with a code sample, if you would be so kind as to relate it to one of the sample packages provided with SSIS so I can run it, that would be great.
I'm having a problem with problems with some custom components I built. I have a custom Connection Manager and Data Flow Source that I have built.
My local PC is 64-bit, but I followed the MS instructions on building and deploying custom components. They run fine inside of VS2005, as long as I have 'Run64BitRuntime" set to 'False'. When I try to run packages using the custom components in 64-bit mode, I get:
Error: 0xC0014005 at : The connection type "<MyConnectionManagerType>" specified for connection manager "PRI" is not recognized as a valid connection manager type. This error is returned when an attempt is made to create a connection manager for an unknown connection type. Check the spelling in the connection type name.
I suspect I need to either change the way I build or deploy the custom components. Can someone shed some light on this?
Hi, As a hobby, I am creating custom SQL Server Integration Services components, which are installed as an add-on to out of the box components. They are mainly used to reduce the development time and minimize the maintance overhead. I would like to know, how I should market them? What are the mediums to use? Since this is the first time, I am creating a software for mass markets, I would really appreciate your help while marketing this baby.
Just FYI, the first commercially available package that I develop is for creating text based files. My package connects to any OLE DB datasource. Executes the query and stores the output in a text file. It is very useful while developing database driven custom feeds. Instead of working with multiple dataflow tasks, data flow sources, you just need to configure one single component and the rest is taken care for you. Since, the package implements the standard Microsoft Task Editor, it is very easy to learn.
Is it possible to make use of VB.Net for Applications and the VBA editor from your own Custom Transform Components (like the script component does)?
I would like to create a component that generates boiler-plate VBA code and that then allows developers to add their own specialised code in certain methods using the editor (just like the script component)
If so, is there any documentation anywhere on how to go about this?
Is it possible to use versioning at all with SSIS components?
My experience so far is that I have not figured out how version them at all, because if I change the version number, then apparently the <property name="UserComponentTypeName" ...> has to be changed in every single reference to the component, in every dataflow, in every single dtsx file that uses the component at all -- this will be a tremendous headache as some of this stuff is in source control. It seemsl clearly infeasible to check out and substitute and change all these references, and check them all back in, except very infrequently -- surely not feasible to do it for each small change.
I am in the habit of versioning dlls and assemblies, and actually changing at least a minor build number for each actual change, so I am a bit unhappy to find that apparently I cannot feasibly do this for a custom SSIS component :(
(I just find a dislike at apparently being forced to leave the version number invariant when I make code changes.)
What I would be very happy with, is if there were a way to adjust all the references in all the dtsx files, to refer to only the major and minor build numbers say, specifying that I am not picky about the revision numbers below those, so that it will find the latest component in the GAC matching the major and minor numbers.
That way I could still update the component version with each code change, which would make me happy, and not have to edit a zillion dtsx occurrences (not having to do that would also make me happy).
Anyone know if that is possible?
I'm hoping it is, and that it is only my ignorance at component references preventing me from doing so.
I've been trying to find a solution to adding javascript and/or HTML to a report (not reportviewer). After a search I found that you can add javascript to a textbox under action properties and it works fine. Is there a way to add custom code anywhere on the page.
The problem is the charts that come with SRS is limited so I'd like to add my own. I'd also like to add some ajax functionality as well. I know it's possible because a company called Dundas is doing it but not exactly sure how they implemented it. We considered Dundas but they are just too expensive for the solution we need.
Is there a way to use a custom component programmatically (I'm using C# to create and run my packages) without installing a compiled .dll of the class into the SQL folders and the GAC? For instance, can I instantiate my custom component and somehow pass it to be used in a task in my code.
Would anyone be aware of anyplace I could find some good information on creating DTS custom tasks? I've come across a couple of articles from SQL Server Magazine, but nothing too substantive... Better yet, if anyone has any success (great or small), I'd like to hear from you and hear some of the things you did, what your custom task does, difficulty. I'm just trying to get an idea of how much work I have ahead of me....
I'm wondering if anyone's accomplished this before - I've been unable to find a whiff of info on how to do this so far.
I'm creating a custom component that I'd like to give a "Derived Column" type of ability to. By that, I mean I'd like to populate a property of my component with an expression (including references to input columns, package variables and functions) and be able to evaluate it at runtime - per row processed by the component.
I would also appreciate any information as to how to provide the interface to allow the user to build such an expression as well - is there a UI function in SSIS I can call to pop up the "expression builder"?
I have a SQL Server 2005 Package (32 bit, 4 processors/16 GB of RAM, 3 GB switch for SSIS) with 1 DataFlow and 5 Custom Components:
> An OLEDB Source that reads a stoed proc from Sybase > 3 Transform Components that perform: 6-required fields check, 4-lookups (3 cached, 1 non-cached), and a calculation that is performed(called) from a .NET Library. > A SQL Destination that uses the .NET BulkCopy Class to insert the data.
It takes about 3 hours to load 1 million records that contain 46 columns (the largest being a varchar(100)). I have tried a good number of tuning options as well as indexing/SQL techniques to improve perfromance.
Some pieces of info: 1) We dynamically create the buffer during runtime based on an SSIS variable. 2) The actions of the components are determined at runtime through config files. We do these two things so we can have multiple data flows run through the same package for better scalability and flexibility. Since we are dynamically creating the buffer I can NOT use any of the out of the box components because of lineage and metadata inconsistency issues.
3) I noticed on PerfMon that I am spooling buffers to disk after 100K rows. I realize this is bad, but what are my alternatives? Should I go to 64 bit? Is there anything else I can do?
I have developed a custom task where I have a lot of properties on. Then I realized that I was missing some properties and added them to my task.
The problem is that my previous version of the task does not "upgrade" properly resulting in me needing to type in all the property values once again. I have severel task that need to upgrade but I can't figure out how to use the functions: Update, CanUpdate and Version
All aspects of my custom task work well except for the icon!
I have included the .ico in the assembly as an "Embedded Resource" and the icon has both 16X16 and 32X32 images. I have edited the IconResource property of the DtsTask attribute as follows:
I've been having an issue with the integration of a third-party DLL into a custom data flow component.
The company sent me a C# project that generates a simple console application. The project includes a class that calls their DLL with DllImport. The console application runs fine.
I created a stand-alone class using the C# class they sent to expose the methods of their DLL. In my custom component, I'm referencing this class to pass data to and from their DLL.
The first method from that stand-alone class that my component encounters simply gets their installation path from the registry and does not use DllImport. That path retrieval works fine. The next method calls a function that is declared with DllImport. Each time the call fails with "System.DllNotFoundException = {"Unable to load DLL AMZip.dll': Exception from HRESULT: 0xE06D7363"}".
I've copied this DLL to countless locations (e.g., the PipelineComponents directory, the project/solution bin directory) and included these paths in all manner of path variables.
What am I missing here? Their DLL is not strong named (does this matter since I'm using DllImport?), my stand-alone class is, and of course, the custom component itself is. I appreciate the help.
I can't find information on how you are suppose to handle the TransactionOption setting inside of a custom task.
I have a custom task that I have developed, and it basically calls a COM+ object that writes data to a database. When I have the task inside of a container that has the transactionoption set to required, and my custom task is set to supported, if one of the X items fail to execute in my custom task I am telling my task to fail the parent, which I thought would rollback everything. But it does not.
Is there someplace that I need to write rollback code in custom SSIS tasks? If there is I can't find any mention of it anywhere. Any examples out there on how to build custom SSIS tasks that support the TransactionOption parameter?
Can anyone tell me where I might find the Class and Sequence Diagrams for the SSIS framework (for Custom Components)?
I've just started trying to create some Custom Transform Components and I'm really struggling to get my head around the component lifecycle (i.e what methods are called when, with what arguments, and why) with just the BOL documentation to guide me.
Ive been using SSIS for a month or two and now find I need to create some custom tasks to perform some performance logging. in the the overloaded ProviderComponentProperties section I am trying to create a property which has the same look as the Expressions properties you find elsewhere (Little + on the left and a group of sub properties when expanded).
Ive have played with creating a IDTSCustomPropertyCollection90 collection then adding my sub properties to it but I cant seem to then add my new collection to the ComponentMetaData.CustomPropertiesCollection.
Im assuming the Expressions parameter is a collection added to the properties collection but I cant figure out how. Any help would be much appreciated.
As part of my SSIS package, a list of sites is created that need to be created on a remote machine. let's say 1000 sites. I need to pass this list to a web service so web service sitting on that machine creates these sites for me. MY SSIS package does not run frequently so I can sacrifice time a little bit to get better functionality.
I need to move the sites that are not created (for any reasons) by web service to another table and successfully created sites to another table, so I need to get confirmation for each site from the web service.
Which option is better?
1) Calling web service for every single record (site) and get the confirmation and then based on the confirmation I move the records accordingly. I know this might be very time consuming, but as I said my SSIS package might only run every six months
OR
2) Sending records to a web service in a batch and get the result. I don€™t know how to do this though.
How do I retrieve the connections (connection managers) collections from Custom Data Flow destination? ComponentMetadata.RuntimeConnectionCollection is empty. I would like to be able to access all the connections defined in the package from the custom data flow task.
I came across code in which it was possible to access the Connections collection using the IDtsConnectionService for custom task (destination). The custom task has access to serviceProvider, whcih can be used to get access to the IDtsConnectionService interface but not the custom data flow task.
I am not sure if this is a correct forum to discuss on the document posted @ http://www.microsoft.com/downloads/details.aspx?familyid=1c2a7dd2-3ec3-4641-9407-a5a337bea7d3&displaylang=en on SQL Server Integration Services (SSIS) Hands on Training - Creating Custom Components.
I am assuming Microsoft Developers are constantly monitoring this forum.
In the document - SSIS Creating a Custom Transformation Component .doc on Page 2 - Exercise 1 - Writing the no-op data flow transformation component - Task 1 - Create a new C# Class Library Project
The textual description talks about creating a new Visual C# Class Library project in VS 2005 but the screenshot accompanying it shows the creation of new "Integration Service Project" in VS 2005.
Please change the screenshot appropriately to avoid confusions.
In my case I have to log the errors raised by any task in a package to either windows event log, text file or SQL server. Also I need to send an email notifications to a group of people telling them about the error.
Now can I use SSIS package logging for logging the errors into the required destinations. I mean right clicking on the package and selecting Logging, then adding the required log providers and enabling the events for logging into those. I think I have to upfront select the log providers to log the error, I will not have the liberty to log the error to the destination, the name of which is passed as a variable to the package. This is okay with me though.
Now what will a custom log provider help me to do in this case. Also can I somehow configure my package to call the send mail task everytime an error is raised.
Also, one more option can be developing a package that only does the error handling. It will take in the paramters or the error codes and descriptions, the destination to write to and a flag to send mail or not for that particular type of error.
In another thread Jamie Thomson very informatively said "The components in SSIS are deliberately atomic (i.e. they do something very specific) so that its easy to put them together to build something greater than the sum of the parts". Which does make a lot of sense. However, I've been finding that I end up having to create exactly the same "pattern" of combined transform components again and again in order to solve the same problem but in different dataflows (or even within the same dataflow). Cut-and-paste-tastic! In order to obtain real re-use, it seems to me like SSIS is crying out for an easy way to create new components by using composition - i.e. the ability to take commonly-used combinations of existing components and create new "super" components (without having to write Custom Transform Components in C#/VB.Net and handle everything in code).
Does anyone know if this sort of functionality is likely to make it into SSIS in the forseeable future?
I am creating a custom transformation component, and a custom user interface for that component.
In my custom UI, I want to show the custom properties, and allow users to edit these properties similar to how the advanced editor shows the properties.
I know in my UI I need to create a "Property Grid". In the properties of this grid, I can select the object I want to display data for, however, the only objects that appear are the objects that I have already created within this UI, and not the actual component object with the custom properties.
How do I go about getting the properties for my transformation component listed in this property grid?