Page 1 of 1

need a logic in parallel job design

Posted: Tue Dec 04, 2007 12:47 am
by neena
cust_id= unique id
my data flow in job :
2 oracle data base ->join stage ->xfm ->MQ serises plug-in

I am reading two tables and joining by using cust_id and sending data to transformer.my source table1 contains cust_id
as key and in 2nd table the keys are cust_id and cust_transx_number

Each cust_id they are spliting:::

<cust_id=1> |
<dtl1> |
. |
. |
. |
</dtl1> |
<dtl2> |
. | ====> Message 1
. |
. |
</dtl2> |
<dtl3> |
. |
. |
. |
</dtl3>
</cust_id>

<cust_id=2> |
<dtl1> |
. |
. |
. |
</dtl1> |
<dtl2> |
. | ====> Message 2
. |
. |
</dtl2> |
<dtl3> |
. |
. |
. |
</dtl3>
</cust_id>


The output should expecting is :::

cust_dtl::

cust_id cust_tranx_number Upc_id
1 1 1000
1 2 1200
1 3 1500

cust_dtl::

2 1 1900
2 2 2200
2 3 2500

any help will be appreciated.

Posted: Tue Dec 04, 2007 5:50 am
by ray.wurlod
Surely it's just a regular join on cust_id? Do this without the Join stage; do it in the Oracle stage.

Posted: Tue Dec 04, 2007 6:03 am
by swapnilverma
Do SQL join in stead of Join Stage.

Any specific reason you used join stage ?
If any then do sorting on keys before joining ....

If you want to track the rejects... then istead if join do Merge and make "Update link" The Reject link( from which you want to get the rejects...)

Hope it helps...

Posted: Tue Dec 04, 2007 6:07 am
by ray.wurlod
A join does not have any rejects. A row is either in the join or it isn't.

Posted: Wed Dec 05, 2007 4:37 pm
by neena
thanks for all, i am using datastage 8.0 here which plugin is better to use here i have websphere MQ stage & websphere MQ connector stage.