Problem in setting a job parameter

[parag.s.27]

Hello All,

I am having a requirement where i am getting a source flat file, in which there is a column which specifies the Material Type code.

Now i have to generate a new target file for each type of material type code.

So what i thought was that, if i can some how change the target file name which is given as a job parameter, at run time in a transformer then whenever the parameter is set the new file name will be there and hence a new target file will be generated.

But the problem is i am not able to set the parameter. Its giving errcode as -1. i am using it as.

i wrote a routine SETPARAM as

Code: Select all

$INCLUDE DSINCLUDE JOBCONTROL.H 
ErrCode = DSSetParam (Jobhandle,ParamName,ParamValue);
Ans = ErrCode

then i called this routine in transformer as : -
SETPARAM(DSJ.ME,"TargetFileName","Value").

Here TargetFileName is the parameter name for my target file.

So i am getting the errcode as -1. Which i assume as of for BADHANDLE. Can anyone please suggest how can i go about it[/code]

[I_Server_Whale]

Is my understanding of your requirement right?


Source Flat File

Code: Select all

col_1          col_2       col_3         Material_Type_Code

1                  A             P                Metal
2                  B             Q                Cloth
3                  C             R                Leather
4                  D             S                Metal
5                  E             T                Stone
6                  F             U                Stone

If you input file is something similar to above. How are your expecting your output apart from the Material_Type_Code going into the file_name?

Whale.

[sb_akarmarkar]

Hi,


It is possible to divide file into multiple file with combination aggregator stage and folder stage....


Thanks,
Anupam

[parag.s.27]

Is my understanding of your requirement right?


Source Flat File

Code: Select all

col_1          col_2       col_3         Material_Type_Code

1                  A             P                Metal
2                  B             Q                Cloth
3                  C             R                Leather
4                  D             S                Metal
5                  E             T                Stone
6                  F             U                Stone

If you input file is something similar to above. How are your expecting your output apart from the Material_Type_Code going into the file_name?

Whale.

I am setting the parameter as

Code: Select all

SETPARAM(DSJ.ME, TargetFileName,source.Material_Type_Code)

hence this will be called for every unique material_type_code for which a different code exists.

[parag.s.27]

Hi,


It is possible to divide file into multiple file with combination aggregator stage and folder stage....


Thanks,
Anupam

Can you please specify How??

Thanks & regards
Parag

[sb_akarmarkar]

Parag,

If data is huge we cant work with this. Better to create two jobs First jobs reads distinct of material code from file (use aggregator to get distinct) pass this value as parameter to second job use DSRunjob routine to trigger second job in first job transformer.
But need to think of performance....


Thanks,
Anupam

[khaja.arshad]

Hi Anupam

Can You explain agai wt u r trying with varaibles V1,V2


Thanks

[parag.s.27]

Hi Anupam

Can You explain agai wt u r trying with varaibles V1,V2


Thanks

Hi Arshad,

What Anupam wants to say is as follows.

Code: Select all

StgOldKey = StgNewKey
StgNewKey = Key

StageVar2 = StageVar3
StageVar1 = concate all columns in this var. as col1:col2...and so on
StageVar3 = if StgNewKey = StgOldKey Then StgRecord2:Char(10):StageVar1 Else StageVar1

and now finally print the value of StageVar2

[chulett]

So what i thought was that, if i can some how change the target file name which is given as a job parameter, at run time in a transformer then whenever the parameter is set the new file name will be there and hence a new target file will be generated.

Just in case it was not clear in the previous posts - this cannot be done on a number of levels. Not only can you not change job parameter values on the fly inside a running job, the Sequential File stage does not support dynamically changing the output filename during a job run in any fashion.

As noted, a Folder stage will allegedly handle that although I've never had a need to actually test or play with it as a target yet. The only stage I'm aware of that has that specific functionality is the XML Output stage. This can be used to create non-XML files with pass-through columns as long as you aren't all that picky about the names it chooses.

For another approach for this, some sort of 'post-processing' to split a single output file up into appropriate parts could be done. You could even handle this in a after-job sub routine, looping through a read of the first file and conditionally closing and opening new output files.

[ray.wurlod]

The original -1 (return status from DSSetParam) is DSJE.BADHANDLE. You had not successfully called DSAttachJob with JobHandle as the handle variable.

[parag.s.27]

So what i thought was that, if i can some how change the target file name which is given as a job parameter, at run time in a transformer then whenever the parameter is set the new file name will be there and hence a new target file will be generated.

Just in case it was not clear in the previous posts - this cannot be done on a number of levels. Not only can you not change job parameter values on the fly inside a running job, the Sequential File stage does not support dynamically changing the output filename during a job run in any fashion.

As noted, a Folder stage will allegedly handle that although I've never had a need to actually test or play with it as a target yet. The only stage I'm aware of that has that specific functionality is the XML Output stage. This can be used to create non-XML files with pass-through columns as long as you aren't all that picky about the names it chooses.

For another approach for this, some sort of 'post-processing' to split a single output file up into appropriate parts could be done. You could even handle this in a after-job sub routine, looping through a read of the first file and conditionally closing and opening new output files.

Hey thanks a lot to all.

This is finally working as per our need. We had to create a new file everytime a new header comes in the material source data.

We are having a structure as Header data then its corresponding multiple detail records and so on.

So using the folder stage its working great