DSXchange

You can very well use job parameters in the column derivation part in a transformer stage. Keep the generated expression in a file and while calling the job, just pass the expressions as job parameters. Use them in your job based on the need.

Vignesh.

You have enough details in Parallel job developer guide. Anyway, lookups have to be done only if the data to be looked up is very small. In that case, lookups' performance will be better than counter parts. On the other hand, if the data volume is high, you have to go for join or merge, depending on...

Your question is not very clear in the first place. But as far I understood, I tell you something which you have to check if useful or not. You generate whatever the where condition you want to generate from the input data in a seperate job and write it into an OS file. Then in the second job, you r...

You can use simple unix commands to cut the first line out of the file. That is the easiest way for doing this !!

So Ray, You mean to say something like if a job with parallel transformer is compiled for a 8 processing nodes, that job will run only in 8 nodes even if the PX actually runs in 16 node production environment. Am I making sense ??

Hi My job uses a funnel stage. The output of funnel is fed to a filter stage where it directs to various tables based on the value of a field. The problem I encountered here is like this. The volume of my source tables are in the order of 20 million rows. My extraction query fetches the rows within ...

Ok Raju, Amitava, Eric and Mandyli. I go with what you guys are saying. But this gives way to one more question. Why does a parallel transformer needs a value in APT Config file parameter to get compiled ?? Can you guys throw some more light on this ??? Thanks in advance !!

Vignesh.

Hi Badri A simple NullToZero() or NullToValue() function will serve your need. As Amos said, you may not get the output what you need on null comparison. So convert both the nulls to some other value before actual lookup takes place. Here care should be taken that you should not convert the null int...

Still the question remains. If there is a situation where either transformer or filter can be used, which one to choose ?? I came across a document stating that using transformer is expensive. If so, in what way is it expensive ?? Ray, Craig, Burney, Bland, Duke and other DataStage Gurus, can you gi...

You can achieve this very easily using stage variables. Keep accumulating the rows into a string without passing them onto the target and after the all the rows are appended pass the string at a stretch to the target.

Hi We are having three different versions of APT Config file and we are passing different file name as a parameter to different job based on the job complexity. Here the issue is, whenever I try to compile the job having a parallel transformer, without giving default value in $APT_CONFIG_FILE parame...

Hi In our project we are using DataStage environment variables for parameterising the Database server name, schema name, log file path. Is it the right way of parameterising these variables or is there any other better way to achieve this ?? If we are changing the value of an environmental variable ...

In simple terms, you want a left outer join to be done on your data. Instead of lookup you use a join stage and give join type as left outer. You will get the required output.

I think the problem with your job is defining metadata. You could have put a different length than actually the values of the field are. Or you might have put a nullable column to non-nullable in metadata. So those records will be merely rejected at the source level itself. Just have a look at your ...

Just to add on to what Ray had said, you can use post job subroutine in which you can count the number of records in the file. If it is equal to the number of rows got from the table (can be drawn from link info) you can issue a truncate command to the table either from the routine itself (for which...