DSXchange

We are interested in Quick runtime and least resource utilization.
There are different look up tables for diff jobs.
Few tables has millions of records others has very few records for look up.
Source data is always 1 million or so.

Hi,
We need to look up tables from Oracle, there are two methods -
1. Either we should use direct database table to look up and fetch
2. Load the look up data set files from table and Look up on these files to fetch.

Which one is better?

We are plaaning to handle only second type of rejects - i.e. Rejects unhandled by DS.

The Business rejects records is out of my project scope, as it involves lot of user defined code.

Ohhh Yeah.... Thats' true...
But I am expecting such a rejection will also get captured automatically as long as I have mentioned correct data types for target.

In 'Px', except transformer stage which are the
other stages where we can use reject link?
Merge, Lookup etc..?

Hi, I need to assign output file name at runtime(with date at postfix). In this forum with 'search' I found solution as below - Use of, ProcessName_#DSJobStartDate#_#DSJobStartTime#.txt in Sequestial File stage. It is applicable for DS Server. Will the same be applicable for Parallel? Or for Paralle...

Yeah, from next time onwards I'll post questions in PX.

So does it mean, if I want to implement Error Handling (reject records) at each process (Extract, Transform & Load), it is not possible.
I can handle only during load?

Thanks for answering, few more questions -
Does it capture whole source row or only some referrence?
Does it capable of giving the reason of rejection?

what is NDM? I can understand loading in flat file from DB2. But from DB2 to IMS how can I go?
And Flat File structure needs to be Like Mainframe? exactly how can I get that structure?

Hi,
In our project we want to implment reject handling.
Suppose there are 100 records at source & only 95 records moved to target. How can I capture those 5 records? The reason behind rejection might be anything.
Along with that, can I generate reason message for each record?

Hi,
Using datastage can I to ETL from DB2 to IMS?
I know I can surely extract from DB2.
So the left part is, can I load in IMS using DataStage?

Nope yet.. but I have discussed this topic with senior architects here as well.. and they are saying we don't need such handling if our target is allowing us Null population.
But surely I'll talk with the concerned team, just to verify.

Nope.. The second question wasn't for you... I need to get it clarify from my team.. they are continuously asking me, how are we handling nullability in the project... and we are not handling in any way b'coz all target fileds, possible of getting null value, are nullable. And from your answer i got...

Thanks,
Then why there is big issue in 'Nullability' handling in my team, I need to get it clarify...

Hi,

There are few fields having nullable in source.
In target I have those fields nullable too. Do I need to handel those fileds, if my source value is null, in Parallel Extender?