Link Partioner Round Robin not happening

Post questions here relative to DataStage Server Edition for such areas as Server job design, DS Basic, Routines, Job Sequences, etc.

Moderators: chulett, rschirm, roy

Post Reply
RShiva
Participant
Posts: 6
Joined: Tue Dec 27, 2005 3:54 am

Link Partioner Round Robin not happening

Post by RShiva »

From a input fixed width Sequential file we are loading to DB2 table where all mapping columns to DB2 table are Key columns with Char format. Job designed as
Seqfile → Tfm→ Link Partioner(3 links)→Tfm→ Db2 table
and in Tfm it is direct mapping no transformation.
Round robin algorithm in Link partioner is used and with no Partioning key. The job gives mismatch in linkcounts statistics into and out from first Tfm stage while writing to Link Partioner stage ( with negative row counts), finishes with Duplicates warning which is not true with data. The round robin is not happening properly. same observation is not seen if we use a dummy sequential file in between tfm and LnkPtrn with same source and target.

thanks
Shiva
ray.wurlod
Participant
Posts: 54607
Joined: Wed Oct 23, 2002 10:52 pm
Location: Sydney, Australia
Contact:

Post by ray.wurlod »

Do you see the same problem with any of the other partitioning algorithms? This is a very surprising finding - Round Robin is about as simple as you can get! Can you please post the actual row counts from the output links of the Link Partitioner stage? And verify that you have row buffering enabled, with what settings.
IBM Software Services Group
Any contribution to this forum is my own opinion and does not necessarily reflect any position that IBM may hold.
RShiva
Participant
Posts: 6
Joined: Tue Dec 27, 2005 3:54 am

Post by RShiva »

Row count for input to tfm is 2368 and out from tfm into link partn is
-12676 and job finishes with duplicates, only 1328 records are loaded to target db2 tables.

Again if I remove the Link Partioner and directly load then it loads all 2368 records .. that is no issue with the source data..
using Lnk Partn and then loading via Tfm here to improve the performance .. that it writes to target table fast by 3 instances at a time which worked fine in many other jobs
I could not understand why it is giving negative row counts....

Thanks
Shiva
ray.wurlod
Participant
Posts: 54607
Joined: Wed Oct 23, 2002 10:52 pm
Location: Sydney, Australia
Contact:

Post by ray.wurlod »

Sounds like a bug. Unsigned integers are sometimes reported as negative numbers, but never in the range about which you are talking. What did your support provider suggest?
IBM Software Services Group
Any contribution to this forum is my own opinion and does not necessarily reflect any position that IBM may hold.
RShiva
Participant
Posts: 6
Joined: Tue Dec 27, 2005 3:54 am

Post by RShiva »

Many Thanks Ray ,
we will forward this to support provider as a bug ,

And our source fields having all fields as keys in Chars , but the negative link count on out put of Tfm link looks a starnge .
Shiva
ray.wurlod
Participant
Posts: 54607
Joined: Wed Oct 23, 2002 10:52 pm
Location: Sydney, Australia
Contact:

Post by ray.wurlod »

:( Unable to reproduce this problem using 7.5.1 on Windows 2003 Server. Round robin partitioning went fine.
IBM Software Services Group
Any contribution to this forum is my own opinion and does not necessarily reflect any position that IBM may hold.
Post Reply