Bug?

When I have two connection managers in the DFT and OLEDB source and destination component.

Associating the different conneciton managers to source and destination seem to go through fine with the code below, but the package generated does not have the connection manager associated with the destination.

I get this error on the destionation on opening up the package in the designer.

<Error> A connection Manager has not been assigned to the runtime connection (24) </Error>

SourceOrDestinationComponent.RuntimeConnectionCollection[0].ConnectionManager = DtsConvert.ToConnectionManager90(pkg.Connections[ConnMgrName]);

However when I assign the ID using the code below it seem to work fine. The documentation doesn't mention needing this step, so is this a bug or a documentation issue?

SourceOrDestinationComponent.RuntimeConnectionCollection[0].ConnectionManagerID = pkg.Connections[ConnMgrName].ID;

Also is the BOL documentation being updated for how to program/configure various tasks in SSIS? Without which it is extremely time consuming with a trial and error method if at all impossible to use them (programmatically)

-tim

Tim,

That step is definately needed and the documentation should say so. If it doesn't say it then use the Send feedback link to tell them. You may be surprised at how effective this is.

If you need anything extra in BOL then ask them for it. It is constantly being updated.

-Jamie

|||

Thanks Jamie, I know the folks working on BOL content appreciate the comments.

This seems like a great opportunity to point out the recent/July refresh of BOL

>>>>>>>>>>>>>>>>>>>>>>>>>>>

Download an updated version of the documentation and tutorials for Microsoft SQL Server 2005. See the "Additional Information" section for an update on SQL Server Express Books Online.

http://www.microsoft.com/downloads/details.aspx?FamilyId=BE6A2C5D-00DF-4220-B133-29C1E0B6585F&displaylang=en

There is a separate download for the sample packages, programs, and databases, at http://www.microsoft.com/downloads/details.aspx?familyid=E719ECF7-9F46-4312-AF89-6AD8702E4E6E&displaylang=en.

|||

Thank you both. The new documentation still does not include the required step.

Craig, BOL seem to have provided clearified some of the existing samples , but it still does not contain samples to how to programmatically most components eg: SCD. Can we expect more samples to be updated to BOL going forward?

Till that time what is the recommended way for people to figure out how to use these apart from trial and error

|||

We are continually updating BOL and I will pass along a note on this one and as Jamie noted, using the feedback mechanism for BOL is really the best way. I can certainly suggest/ask for doc changes but requests from customers always carry more weight. One complication is that every connection mgr may behave differently. Connections and data providers are tricky as each has its own behavior.

sql

BUG REPORT: ReadWriteVariables different between Script Task and Script Component

I have just spent two hours banging my head off the wall to figure this one out... wondering if I had a problem with variable scope or something...

I had a Script Task which was taking a few package level variables to use in the script.

So I put the list of variables in the ReadWriteVariables property.

Specified them like this:

"User::FileBeingProcessed, User::FileSource"

And all was well....

Then later on decided to use the same set of variables in a Script Component contained in a Data Flow.

And kept kept getting a "..variable not found.." error dialog when I clicked on Design Script.

Finally the solution came to me: Spaces in the list of variables !

In a script task you can have a space after a comma in the list... but in a script component you CAN'T !!

So what works in a Script Task is :

Task : "User::FileBeingProcessed, User::FileSource"

But that will not work in a Script Component...

See the difference?:

Task :"User::FileBeingProcessed, User::FileSource"
Component:"User::FileBeingProcessed,User::FileSource"

I know the BOL says to use a comma separated list.. but.. some consistency would be nice...

PJ

I agree..that's a bit...err...cra@.p!

I always thought both used the syntax without a space in there and hence that what I always use - and hence I never have a problem with it.

Bug it! http://lab.msdn.microsoft.com/productfeedback/default.aspx

-Jamie

|||

Oh yeah, if you worked without the spaces you'd never see the problem... its just if you're an awkward fecker like me... heh heh...

Very misleading tho.. I was thrashing around for ages worried about scope (since the Script component is only down inside a data flow) and then applying service packs.. etc...

Still.. yer eventually glad just to find out yer not going crazy...

heh,

PJ

buffer.SetString not writing values!

Hello all,

I'm writing a custom component similar to the "Derived Column" transformation, in which a value of some input column is "decoded" into an output column.

Here's the code:

Code Snippet

public override void ProcessInput(int inputID, PipelineBuffer buffer){

//TODO

setupTraceListener();

IDTSOutput90 outp = ComponentMetaData.OutputCollection["DecodeOutput"];

int outId = outp.ID;

int errorOut = -1;

IDTSInput90 inp = ComponentMetaData.InputCollection.GetObjectByID(inputID);

if( ComponentMetaData.OutputCollection.Count > 1){errorOut = ComponentMetaData.OutputCollection["DecodeErrors"].ID;

}

string sourceValue;

if(!buffer.EndOfRowset){

while(buffer.NextRow()){

if(!buffer.IsNull(inputColumnBufferIndex)){

switch(inp.InputColumnCollection[0].DataType){

case DataType.DT_WSTR:

sourceValue = buffer.GetString(this.inputColumnBufferIndex);

break;

case DataType.DT_I4:

sourceValue = buffer.GetInt32(this.inputColumnBufferIndex).ToString();

break;

default: throw new Exception("Invalid Data Type!");

}

sourceValue = sourceValue.Trim();

try{

object decodedValueObj = this.mappings[sourceValue];

if(decodedValueObj == null){

throw new Exception("No mapping!");

}

string decodedValue = decodedValueObj.ToString();

switch(outp.OutputColumnCollection[0].DataType){

case DataType.DT_WSTR:

buffer.SetString(this.outputColumnBufferIndex, decodedValue);

break;

case DataType.DT_I4:

buffer.SetInt32(this.outputColumnBufferIndex, int.Parse(decodedValue));

break;

default:

throw new IOException("Invalid Data Type!");

}

buffer.DirectRow(outId);

}catch(IOException fake){

throw fake;

}catch(Exception e){

redirectOrFail(inp, errorOut, outId, buffer, e);

}

}else{

redirectOrFail(inp,errorOut,outId,buffer,new Exception("Null values cannot be mapped"));

}

}

}

}

mappings is a hash map filled in PreExecute by accessing a database, outputColumnBufferIndex is calculated like this:

Code Snippet

IDTSInput90 input = ComponentMetaData.InputCollection[0];

IDTSOutput90 output = ComponentMetaData.OutputCollection["DecodeOutput"];

IDTSInputColumn90 inputColumn = input.InputColumnCollection[0];

IDTSOutputColumn90 outputColumn = output.OutputColumnCollection[0];

this.inputColumnBufferIndex =

BufferManager.FindColumnByLineageID(input.Buffer, inputColumn.LineageID);

this.outputColumnBufferIndex =

BufferManager.FindColumnByLineageID(input.Buffer, outputColumn.LineageID);

I've highlighted the parts that interest me, I've been testing the component thoroughly and it is getting the input data and mappings right, but it won't write the Decoded value to the given column and leaves it empty, even though the decodedValue variable is not null.

I'm guessing theres a problem when I call buffer.SetString() but I haven't the slightest idea of what that could be.

Any help would be greatly appreciated!

Have you debugged it? Do inputColumnBufferIndex and outputColumnBufferIndex values make sense?

Do you only have one input and one output column in your component?

Are you getting any errors from SetString?

Thanks,

Bob

|||

Thank you for your reply Bob,

Yes, I did debug it and the values made sense. Thats how I know that the decodedValue was being also properly initialized.

It's working now by using:

Code Snippet

buffer[this.outputColumnBufferIndex] = decodedValue;

Don't ask me why.

I'm having all sorts of problems that have very easy but completely undocumented solutions like having to override PerformUpgrade to update the UserComponentTypeName attribute and the Version, even though I'm not changing either one as the component is still in development and not in use in any "real" packages.

Maybe my installation is corrupted somehow?

|||

Hmm, you should not be required to implement PerformUpgrade unless you are really changing your component.

Could you check your ProvideComponentProperties implementation and if there is something that might be messing up the settings?

Thanks,

Bob

|||

I don't know what to look for, this is my 3rd week with C# and SSIS.

Here's the basic idea:

Code Snippet

public override void ProvideComponentProperties(){

//FIXME

base.ProvideComponentProperties();

(set the name, dispositions, etc...)

base.RemoveAllInputsOutputsAndCustomProperties();

(add synchronous inputs and outputs)

(add custom properties)

}

I suspect it could be I'm calling RemoveAllInpusOutputsAndCustomProperties(); after base.ProvideComponentProperties(), but I don't want to touch it just yet. What are your thoughts?

Thank you very much again.

|||

Yeah, that is it.

You can safely remove that line as ProvideComponentProperties is called only once in the SSIS designer.

Alternatively, you may move that line before base.ProvideComponentProperties().

Thanks,

Bob

|||Thank you very much for your help|||

Adrian, this is not related to your question, but there is a bug in how buffer.EndOfRowset is used in the ProcessInput method in the first code sample, please see my blog for details

http://blogs.msdn.com/michen/archive/2007/08/31/Buffer.EndOfRowset.aspx

|||

Thank you very much for that link! I'll be changing that right away.

buffer.SetString not writing values!

Hello all,

I'm writing a custom component similar to the "Derived Column" transformation, in which a value of some input column is "decoded" into an output column.

Here's the code:

Code Snippet

public override void ProcessInput(int inputID, PipelineBuffer buffer){

//TODO

setupTraceListener();

IDTSOutput90 outp = ComponentMetaData.OutputCollection["DecodeOutput"];

int outId = outp.ID;

int errorOut = -1;

IDTSInput90 inp = ComponentMetaData.InputCollection.GetObjectByID(inputID);

if( ComponentMetaData.OutputCollection.Count > 1){errorOut = ComponentMetaData.OutputCollection["DecodeErrors"].ID;

}

string sourceValue;

if(!buffer.EndOfRowset){

while(buffer.NextRow()){

if(!buffer.IsNull(inputColumnBufferIndex)){

switch(inp.InputColumnCollection[0].DataType){

case DataType.DT_WSTR:

sourceValue = buffer.GetString(this.inputColumnBufferIndex);

break;

case DataType.DT_I4:

sourceValue = buffer.GetInt32(this.inputColumnBufferIndex).ToString();

break;

default: throw new Exception("Invalid Data Type!");

}

sourceValue = sourceValue.Trim();

try{

object decodedValueObj = this.mappings[sourceValue];

if(decodedValueObj == null){

throw new Exception("No mapping!");

}

string decodedValue = decodedValueObj.ToString();

switch(outp.OutputColumnCollection[0].DataType){

case DataType.DT_WSTR:

buffer.SetString(this.outputColumnBufferIndex, decodedValue);

break;

case DataType.DT_I4:

buffer.SetInt32(this.outputColumnBufferIndex, int.Parse(decodedValue));

break;

default:

throw new IOException("Invalid Data Type!");

}

buffer.DirectRow(outId);

}catch(IOException fake){

throw fake;

}catch(Exception e){

redirectOrFail(inp, errorOut, outId, buffer, e);

}

}else{

redirectOrFail(inp,errorOut,outId,buffer,new Exception("Null values cannot be mapped"));

}

}

}

}

mappings is a hash map filled in PreExecute by accessing a database, outputColumnBufferIndex is calculated like this:

Code Snippet

IDTSInput90 input = ComponentMetaData.InputCollection[0];

IDTSOutput90 output = ComponentMetaData.OutputCollection["DecodeOutput"];

IDTSInputColumn90 inputColumn = input.InputColumnCollection[0];

IDTSOutputColumn90 outputColumn = output.OutputColumnCollection[0];

this.inputColumnBufferIndex =

BufferManager.FindColumnByLineageID(input.Buffer, inputColumn.LineageID);

this.outputColumnBufferIndex =

BufferManager.FindColumnByLineageID(input.Buffer, outputColumn.LineageID);

I've highlighted the parts that interest me, I've been testing the component thoroughly and it is getting the input data and mappings right, but it won't write the Decoded value to the given column and leaves it empty, even though the decodedValue variable is not null.

I'm guessing theres a problem when I call buffer.SetString() but I haven't the slightest idea of what that could be.

Any help would be greatly appreciated!

Have you debugged it? Do inputColumnBufferIndex and outputColumnBufferIndex values make sense?

Do you only have one input and one output column in your component?

Are you getting any errors from SetString?

Thanks,

Bob

|||

Thank you for your reply Bob,

Yes, I did debug it and the values made sense. Thats how I know that the decodedValue was being also properly initialized.

It's working now by using:

Code Snippet

buffer[this.outputColumnBufferIndex] = decodedValue;

Don't ask me why.

I'm having all sorts of problems that have very easy but completely undocumented solutions like having to override PerformUpgrade to update the UserComponentTypeName attribute and the Version, even though I'm not changing either one as the component is still in development and not in use in any "real" packages.

Maybe my installation is corrupted somehow?

|||

Hmm, you should not be required to implement PerformUpgrade unless you are really changing your component.

Could you check your ProvideComponentProperties implementation and if there is something that might be messing up the settings?

Thanks,

Bob

|||

I don't know what to look for, this is my 3rd week with C# and SSIS.

Here's the basic idea:

Code Snippet

public override void ProvideComponentProperties(){

//FIXME

base.ProvideComponentProperties();

(set the name, dispositions, etc...)

base.RemoveAllInputsOutputsAndCustomProperties();

(add synchronous inputs and outputs)

(add custom properties)

}

I suspect it could be I'm calling RemoveAllInpusOutputsAndCustomProperties(); after base.ProvideComponentProperties(), but I don't want to touch it just yet. What are your thoughts?

Thank you very much again.

|||

Yeah, that is it.

You can safely remove that line as ProvideComponentProperties is called only once in the SSIS designer.

Alternatively, you may move that line before base.ProvideComponentProperties().

Thanks,

Bob

|||Thank you very much for your help|||

Adrian, this is not related to your question, but there is a bug in how buffer.EndOfRowset is used in the ProcessInput method in the first code sample, please see my blog for details

http://blogs.msdn.com/michen/archive/2007/08/31/Buffer.EndOfRowset.aspx

|||

Thank you very much for that link! I'll be changing that right away.

buffer.SetString not writing values!

Hello all,

I'm writing a custom component similar to the "Derived Column" transformation, in which a value of some input column is "decoded" into an output column.

Here's the code:

Code Snippet

public override void ProcessInput(int inputID, PipelineBuffer buffer){

//TODO

setupTraceListener();

IDTSOutput90 outp = ComponentMetaData.OutputCollection["DecodeOutput"];

int outId = outp.ID;

int errorOut = -1;

IDTSInput90 inp = ComponentMetaData.InputCollection.GetObjectByID(inputID);

if( ComponentMetaData.OutputCollection.Count > 1){errorOut = ComponentMetaData.OutputCollection["DecodeErrors"].ID;

}

string sourceValue;

if(!buffer.EndOfRowset){

while(buffer.NextRow()){

if(!buffer.IsNull(inputColumnBufferIndex)){

switch(inp.InputColumnCollection[0].DataType){

case DataType.DT_WSTR:

sourceValue = buffer.GetString(this.inputColumnBufferIndex);

break;

case DataType.DT_I4:

sourceValue = buffer.GetInt32(this.inputColumnBufferIndex).ToString();

break;

default: throw new Exception("Invalid Data Type!");

}

sourceValue = sourceValue.Trim();

try{

object decodedValueObj = this.mappings[sourceValue];

if(decodedValueObj == null){

throw new Exception("No mapping!");

}

string decodedValue = decodedValueObj.ToString();

switch(outp.OutputColumnCollection[0].DataType){

case DataType.DT_WSTR:

buffer.SetString(this.outputColumnBufferIndex, decodedValue);

break;

case DataType.DT_I4:

buffer.SetInt32(this.outputColumnBufferIndex, int.Parse(decodedValue));

break;

default:

throw new IOException("Invalid Data Type!");

}

buffer.DirectRow(outId);

}catch(IOException fake){

throw fake;

}catch(Exception e){

redirectOrFail(inp, errorOut, outId, buffer, e);

}

}else{

redirectOrFail(inp,errorOut,outId,buffer,new Exception("Null values cannot be mapped"));

}

}

}

}

mappings is a hash map filled in PreExecute by accessing a database, outputColumnBufferIndex is calculated like this:

Code Snippet

IDTSInput90 input = ComponentMetaData.InputCollection[0];

IDTSOutput90 output = ComponentMetaData.OutputCollection["DecodeOutput"];

IDTSInputColumn90 inputColumn = input.InputColumnCollection[0];

IDTSOutputColumn90 outputColumn = output.OutputColumnCollection[0];

this.inputColumnBufferIndex =

BufferManager.FindColumnByLineageID(input.Buffer, inputColumn.LineageID);

this.outputColumnBufferIndex =

BufferManager.FindColumnByLineageID(input.Buffer, outputColumn.LineageID);

I've highlighted the parts that interest me, I've been testing the component thoroughly and it is getting the input data and mappings right, but it won't write the Decoded value to the given column and leaves it empty, even though the decodedValue variable is not null.

I'm guessing theres a problem when I call buffer.SetString() but I haven't the slightest idea of what that could be.

Any help would be greatly appreciated!

Have you debugged it? Do inputColumnBufferIndex and outputColumnBufferIndex values make sense?

Do you only have one input and one output column in your component?

Are you getting any errors from SetString?

Thanks,

Bob

|||

Thank you for your reply Bob,

Yes, I did debug it and the values made sense. Thats how I know that the decodedValue was being also properly initialized.

It's working now by using:

Code Snippet

buffer[this.outputColumnBufferIndex] = decodedValue;

Don't ask me why.

I'm having all sorts of problems that have very easy but completely undocumented solutions like having to override PerformUpgrade to update the UserComponentTypeName attribute and the Version, even though I'm not changing either one as the component is still in development and not in use in any "real" packages.

Maybe my installation is corrupted somehow?

|||

Hmm, you should not be required to implement PerformUpgrade unless you are really changing your component.

Could you check your ProvideComponentProperties implementation and if there is something that might be messing up the settings?

Thanks,

Bob

|||

I don't know what to look for, this is my 3rd week with C# and SSIS.

Here's the basic idea:

Code Snippet

public override void ProvideComponentProperties(){

//FIXME

base.ProvideComponentProperties();

(set the name, dispositions, etc...)

base.RemoveAllInputsOutputsAndCustomProperties();

(add synchronous inputs and outputs)

(add custom properties)

}

I suspect it could be I'm calling RemoveAllInpusOutputsAndCustomProperties(); after base.ProvideComponentProperties(), but I don't want to touch it just yet. What are your thoughts?

Thank you very much again.

|||

Yeah, that is it.

You can safely remove that line as ProvideComponentProperties is called only once in the SSIS designer.

Alternatively, you may move that line before base.ProvideComponentProperties().

Thanks,

Bob

|||Thank you very much for your help|||

Adrian, this is not related to your question, but there is a bug in how buffer.EndOfRowset is used in the ProcessInput method in the first code sample, please see my blog for details

http://blogs.msdn.com/michen/archive/2007/08/31/Buffer.EndOfRowset.aspx

|||

Thank you very much for that link! I'll be changing that right away.

buffer.SetString not writing values!

Hello all,

I'm writing a custom component similar to the "Derived Column" transformation, in which a value of some input column is "decoded" into an output column.

Here's the code:

Code Snippet

public override void ProcessInput(int inputID, PipelineBuffer buffer){

//TODO

setupTraceListener();

IDTSOutput90 outp = ComponentMetaData.OutputCollection["DecodeOutput"];

int outId = outp.ID;

int errorOut = -1;

IDTSInput90 inp = ComponentMetaData.InputCollection.GetObjectByID(inputID);

if( ComponentMetaData.OutputCollection.Count > 1){errorOut = ComponentMetaData.OutputCollection["DecodeErrors"].ID;

}

string sourceValue;

if(!buffer.EndOfRowset){

while(buffer.NextRow()){

if(!buffer.IsNull(inputColumnBufferIndex)){

switch(inp.InputColumnCollection[0].DataType){

case DataType.DT_WSTR:

sourceValue = buffer.GetString(this.inputColumnBufferIndex);

break;

case DataType.DT_I4:

sourceValue = buffer.GetInt32(this.inputColumnBufferIndex).ToString();

break;

default: throw new Exception("Invalid Data Type!");

}

sourceValue = sourceValue.Trim();

try{

object decodedValueObj = this.mappings[sourceValue];

if(decodedValueObj == null){

throw new Exception("No mapping!");

}

string decodedValue = decodedValueObj.ToString();

switch(outp.OutputColumnCollection[0].DataType){

case DataType.DT_WSTR:

buffer.SetString(this.outputColumnBufferIndex, decodedValue);

break;

case DataType.DT_I4:

buffer.SetInt32(this.outputColumnBufferIndex, int.Parse(decodedValue));

break;

default:

throw new IOException("Invalid Data Type!");

}

buffer.DirectRow(outId);

}catch(IOException fake){

throw fake;

}catch(Exception e){

redirectOrFail(inp, errorOut, outId, buffer, e);

}

}else{

redirectOrFail(inp,errorOut,outId,buffer,new Exception("Null values cannot be mapped"));

}

}

}

}

mappings is a hash map filled in PreExecute by accessing a database, outputColumnBufferIndex is calculated like this:

Code Snippet

IDTSInput90 input = ComponentMetaData.InputCollection[0];

IDTSOutput90 output = ComponentMetaData.OutputCollection["DecodeOutput"];

IDTSInputColumn90 inputColumn = input.InputColumnCollection[0];

IDTSOutputColumn90 outputColumn = output.OutputColumnCollection[0];

this.inputColumnBufferIndex =

BufferManager.FindColumnByLineageID(input.Buffer, inputColumn.LineageID);

this.outputColumnBufferIndex =

BufferManager.FindColumnByLineageID(input.Buffer, outputColumn.LineageID);

I've highlighted the parts that interest me, I've been testing the component thoroughly and it is getting the input data and mappings right, but it won't write the Decoded value to the given column and leaves it empty, even though the decodedValue variable is not null.

I'm guessing theres a problem when I call buffer.SetString() but I haven't the slightest idea of what that could be.

Any help would be greatly appreciated!

Have you debugged it? Do inputColumnBufferIndex and outputColumnBufferIndex values make sense?

Do you only have one input and one output column in your component?

Are you getting any errors from SetString?

Thanks,

Bob

|||

Thank you for your reply Bob,

Yes, I did debug it and the values made sense. Thats how I know that the decodedValue was being also properly initialized.

It's working now by using:

Code Snippet

buffer[this.outputColumnBufferIndex] = decodedValue;

Don't ask me why.

I'm having all sorts of problems that have very easy but completely undocumented solutions like having to override PerformUpgrade to update the UserComponentTypeName attribute and the Version, even though I'm not changing either one as the component is still in development and not in use in any "real" packages.

Maybe my installation is corrupted somehow?

|||

Hmm, you should not be required to implement PerformUpgrade unless you are really changing your component.

Could you check your ProvideComponentProperties implementation and if there is something that might be messing up the settings?

Thanks,

Bob

|||

I don't know what to look for, this is my 3rd week with C# and SSIS.

Here's the basic idea:

Code Snippet

public override void ProvideComponentProperties(){

//FIXME

base.ProvideComponentProperties();

(set the name, dispositions, etc...)

base.RemoveAllInputsOutputsAndCustomProperties();

(add synchronous inputs and outputs)

(add custom properties)

}

I suspect it could be I'm calling RemoveAllInpusOutputsAndCustomProperties(); after base.ProvideComponentProperties(), but I don't want to touch it just yet. What are your thoughts?

Thank you very much again.

|||

Yeah, that is it.

You can safely remove that line as ProvideComponentProperties is called only once in the SSIS designer.

Alternatively, you may move that line before base.ProvideComponentProperties().

Thanks,

Bob

|||Thank you very much for your help|||

Adrian, this is not related to your question, but there is a bug in how buffer.EndOfRowset is used in the ProcessInput method in the first code sample, please see my blog for details

http://blogs.msdn.com/michen/archive/2007/08/31/Buffer.EndOfRowset.aspx

|||

Thank you very much for that link! I'll be changing that right away.

buffer.SetString not writing values!

Hello all,

I'm writing a custom component similar to the "Derived Column" transformation, in which a value of some input column is "decoded" into an output column.

Here's the code:

Code Snippet

public override void ProcessInput(int inputID, PipelineBuffer buffer){

//TODO

setupTraceListener();

IDTSOutput90 outp = ComponentMetaData.OutputCollection["DecodeOutput"];

int outId = outp.ID;

int errorOut = -1;

IDTSInput90 inp = ComponentMetaData.InputCollection.GetObjectByID(inputID);

if( ComponentMetaData.OutputCollection.Count > 1){errorOut = ComponentMetaData.OutputCollection["DecodeErrors"].ID;

}

string sourceValue;

if(!buffer.EndOfRowset){

while(buffer.NextRow()){

if(!buffer.IsNull(inputColumnBufferIndex)){

switch(inp.InputColumnCollection[0].DataType){

case DataType.DT_WSTR:

sourceValue = buffer.GetString(this.inputColumnBufferIndex);

break;

case DataType.DT_I4:

sourceValue = buffer.GetInt32(this.inputColumnBufferIndex).ToString();

break;

default: throw new Exception("Invalid Data Type!");

}

sourceValue = sourceValue.Trim();

try{

object decodedValueObj = this.mappings[sourceValue];

if(decodedValueObj == null){

throw new Exception("No mapping!");

}

string decodedValue = decodedValueObj.ToString();

switch(outp.OutputColumnCollection[0].DataType){

case DataType.DT_WSTR:

buffer.SetString(this.outputColumnBufferIndex, decodedValue);

break;

case DataType.DT_I4:

buffer.SetInt32(this.outputColumnBufferIndex, int.Parse(decodedValue));

break;

default:

throw new IOException("Invalid Data Type!");

}

buffer.DirectRow(outId);

}catch(IOException fake){

throw fake;

}catch(Exception e){

redirectOrFail(inp, errorOut, outId, buffer, e);

}

}else{

redirectOrFail(inp,errorOut,outId,buffer,new Exception("Null values cannot be mapped"));

}

}

}

}

mappings is a hash map filled in PreExecute by accessing a database, outputColumnBufferIndex is calculated like this:

Code Snippet

IDTSInput90 input = ComponentMetaData.InputCollection[0];

IDTSOutput90 output = ComponentMetaData.OutputCollection["DecodeOutput"];

IDTSInputColumn90 inputColumn = input.InputColumnCollection[0];

IDTSOutputColumn90 outputColumn = output.OutputColumnCollection[0];

this.inputColumnBufferIndex =

BufferManager.FindColumnByLineageID(input.Buffer, inputColumn.LineageID);

this.outputColumnBufferIndex =

BufferManager.FindColumnByLineageID(input.Buffer, outputColumn.LineageID);

I've highlighted the parts that interest me, I've been testing the component thoroughly and it is getting the input data and mappings right, but it won't write the Decoded value to the given column and leaves it empty, even though the decodedValue variable is not null.

I'm guessing theres a problem when I call buffer.SetString() but I haven't the slightest idea of what that could be.

Any help would be greatly appreciated!

Have you debugged it? Do inputColumnBufferIndex and outputColumnBufferIndex values make sense?

Do you only have one input and one output column in your component?

Are you getting any errors from SetString?

Thanks,

Bob

|||

Thank you for your reply Bob,

Yes, I did debug it and the values made sense. Thats how I know that the decodedValue was being also properly initialized.

It's working now by using:

Code Snippet

buffer[this.outputColumnBufferIndex] = decodedValue;

Don't ask me why.

I'm having all sorts of problems that have very easy but completely undocumented solutions like having to override PerformUpgrade to update the UserComponentTypeName attribute and the Version, even though I'm not changing either one as the component is still in development and not in use in any "real" packages.

Maybe my installation is corrupted somehow?

|||

Hmm, you should not be required to implement PerformUpgrade unless you are really changing your component.

Could you check your ProvideComponentProperties implementation and if there is something that might be messing up the settings?

Thanks,

Bob

|||

I don't know what to look for, this is my 3rd week with C# and SSIS.

Here's the basic idea:

Code Snippet

public override void ProvideComponentProperties(){

//FIXME

base.ProvideComponentProperties();

(set the name, dispositions, etc...)

base.RemoveAllInputsOutputsAndCustomProperties();

(add synchronous inputs and outputs)

(add custom properties)

}

I suspect it could be I'm calling RemoveAllInpusOutputsAndCustomProperties(); after base.ProvideComponentProperties(), but I don't want to touch it just yet. What are your thoughts?

Thank you very much again.

|||

Yeah, that is it.

You can safely remove that line as ProvideComponentProperties is called only once in the SSIS designer.

Alternatively, you may move that line before base.ProvideComponentProperties().

Thanks,

Bob

|||Thank you very much for your help|||

Adrian, this is not related to your question, but there is a bug in how buffer.EndOfRowset is used in the ProcessInput method in the first code sample, please see my blog for details

http://blogs.msdn.com/michen/archive/2007/08/31/Buffer.EndOfRowset.aspx

|||

Thank you very much for that link! I'll be changing that right away.

Breakpoints in script component

Does anybody know if the breakpoints are supported in the script component?

I've got a script that I want to debug and make sure it's doing what it should be but everytime I set a breakpoint it loses them when I click ok on the script editor. I've tried setting precompile to false but thats not working. What do I need to do to be able to use breakpoints?

Saitham8Nope, breakpoints in script components are not supported in this version. Only script tasks support breakpoints.|||why? if breakpoints in script components are disabled how are we supposed to debug our code?|||

Saitham8 wrote:

why?

The usual time/resources limitations - did not have time to implement them.

|||go figure. (somebody really needs to create a work around for this time limitation thing :P ) any chance of it getting put in at a later time? and is there a work around?|||the only way i know how to debug a script component at run-time is with the data viewer and the row count component.

BreakpointManager,BreakpointSite

Hey

i am creating a custom component and want to give the user a breakpoint or 3. in AcceptBreakpointManager I say

this._bpm = BreakpointManager

_bpm.CreateBreakpointTarget(1,"You hit #1");
_bpm.CreateBreakpointTarget(2,"You hit #2");
It was my understanding that I would be able to see these breakpoint descriptions in the list of breakpoints for my custom task. I cannot. All i see is OnCustomEvent which unless I have missed something does not tell me which breakpoint was hit.

In Execute() I check to see if the breakpoint is enabled (checked) and go from there

Did I miss something?

Allan

Allan, this looks right to me. I'm assuming _bmp is valid when you call it? Are you using the breakPointMgr that's getting passed in as a parameter?

K

|||Hey Kirk.

Yeh _bpm is a BreakpointManager at the class level and would seen to be valid. breakpointManager is the value passed in to the method.

Ok So i just figured why it does not work.

I had removed the interface IDTSBreakpointSite from my task. Now i have added this back it works.

Once thing though. In BOL it says that IDTSAcceptBreakpointSite.AcceptBreakpointManager is called by the runtime and not used in code. BOL then goes on to give an example (I have always overriden this method also).

To what does BOL when it says "not used in code" refer?

Thanks Kirk.

Allan

|||"I had removed the interface IDTSBreakpointSite from my task."
Yep, gotta have that or the TaskHost won't forward the calls.

"To what does BOL when it says "not used in code" refer?"
I have no idea what that means. It might mean that the task itself should not call that function, which is true. But, I dunno for sure what, "Not used in code" means.