Use of 'C' in Parallel Jobs
Posted: Thu Dec 22, 2005 9:13 am
I've read a great many threads on Parallel Jobs vs Server Jobs on topics such as which to use and when, BASIC vs 'C', and the future threat to Server Jobs being removed from the Product going forward. These fears have been alleviated somewhat by some of the comments on this forum.
However, it seems that most new developers in my Company are now going down the path of Parallel Jobs regardless of the task in hand - I guess peoples mindsets are changing against Server Jobs and don't want to waste time learning something that may disappear - that is another topic outwith the scope of this question however. I am merely familiarising myself with Parallel Jobs at the moment.
My use of Datastage is more EL than ETL - I only have one transformation in 10 Job Sequences/60 Server Jobs that I have written. All of my Job
Sequences have multiple Server Jobs in them and a number of BASIC routines. I have no before/after routines or any Job Control so any BASIC code that I have written is consigned to Routines in the Job Sequence.
Going forward I may adopt the methodology of Job Sequences to knit Parallel Jobs together (instead of Server Jobs) to carry out EL processing.
With the above in mind the question I have is related to coding routines in BASIC vs 'C'.
(1) If I use the above methodology am I correct in assuming that I will have no need to write 'C' routines - my only routines that I write will be in the Job Sequence and therefore be in BASIC. Is this a fair assumption ? I have not looked at all of the Stages in Parallel Jobs in great detail yet (slowly working through them) and so far haven't come across any that may need 'C' code added to them.
(2) If Server Jobs were to disappear from DataStage in, say 2-3 years, what would happen to Job Sequences. I believed that they were special
types of Server Jobs ?
Apologies if my undertanding of this is incorrect.
However, it seems that most new developers in my Company are now going down the path of Parallel Jobs regardless of the task in hand - I guess peoples mindsets are changing against Server Jobs and don't want to waste time learning something that may disappear - that is another topic outwith the scope of this question however. I am merely familiarising myself with Parallel Jobs at the moment.
My use of Datastage is more EL than ETL - I only have one transformation in 10 Job Sequences/60 Server Jobs that I have written. All of my Job
Sequences have multiple Server Jobs in them and a number of BASIC routines. I have no before/after routines or any Job Control so any BASIC code that I have written is consigned to Routines in the Job Sequence.
Going forward I may adopt the methodology of Job Sequences to knit Parallel Jobs together (instead of Server Jobs) to carry out EL processing.
With the above in mind the question I have is related to coding routines in BASIC vs 'C'.
(1) If I use the above methodology am I correct in assuming that I will have no need to write 'C' routines - my only routines that I write will be in the Job Sequence and therefore be in BASIC. Is this a fair assumption ? I have not looked at all of the Stages in Parallel Jobs in great detail yet (slowly working through them) and so far haven't come across any that may need 'C' code added to them.
(2) If Server Jobs were to disappear from DataStage in, say 2-3 years, what would happen to Job Sequences. I believed that they were special
types of Server Jobs ?
Apologies if my undertanding of this is incorrect.