Controlling the execution of Jobs in Datastage Sequence

[synsog]

Hi Everyone,

We need to design a sequence in such a way that it execute only required jobs and not all attached to that sequence.

Sample Job Flow in Sequence.

Extract --> Transform A1 --> Transform A2 --> Load

Say, Our Sequence is designed to process data for 2 countries A1, A2.

So, when it is running for country A1, it should only execute below flow

Extract --> Transform A1 --> Load **Skipped Transform A2**

And when it is running for country A2, it should only execute below flow

Extract --> Transform A2 --> Load **Skipped Transform A1**

is there a way to achieve this ?

Thanks in Advance!

Regards,
synsog.

[ray.wurlod]

Only ever one country, or possibly both? Either way you set up parameter(s) in the sequence to specify which country/countries, and use trigger expressions - possibly in Nested Condition activities - to specify what is to occur. Is there any dependency (for example country A2 runs only after country A1)? This will affect the design but not the mechanism.

[synsog]

No - There is no such dependency. This sequence will be automated through scheduler -Which ever country file is created first will get the priority. We have a country code parameter at sequence level to differentiate.

We would have created separate sequence for each country. But in future, we ould be adding more countries. Almost all jobs will remain same. We are expecting only 1 02 jobs to be country specific.

I did not undertood - How to put this iin nested condition.

Regards,
synsog

[ray.wurlod]

Code: Select all

Nested Condition  ---->  JobA1
#jpCountry# = "A1"

Nested Condition  ---->  JobA2
#jpCountry# = "A2"

Even if the job is multi-instance it can be called from different Job activities with different invocation IDs - possibly the country.

[prasannakumarkk]

Can be like this

Code: Select all

Extract --> Nested Condition -->Transform A1 
                               --> Transform A2 --> Seqeuncer(All) --> Load
                               --> Transform AN

In the nested condition activity, in the expression you can give the condition of country code. The above design run the jobs in parallel. If the transformation logic is same but only the country code are different, then you can parameterize it and make the transformer job multi instance.

[ray.wurlod]

The specification was that "only one country will be processed at a time".

While that does not invalidate your design, you've over-fulfilled the requirement.

[prasannakumarkk]

Specifying any in sequencer will hold good i guess;