Hi folks,
I have a doubt on oracle enterprise stage's degree of parallelism when it acts as source.
When read method= Table (stage automatically reads in parallel).
what is the degree of parallelism here if my table is partitioned?
what is degree of parallelism here if my table is not partitioned?
when read method = user defined SQL or auto generated SQL
and and partition table property is not set stage reads sequentially. Is it sequentially per partition or runs on only one Node i.e conductor node?
when read method = user defined SQL or auto generated SQL,
and partition table property is set
I think degree of parallelism is no of partitions rather than no of nodes in ATP_CONFIG_FILE . Please confirm this.
Oracle enterprise stage- Degree of parallelism
Moderators: chulett, rschirm, roy
Oracle enterprise stage- Degree of parallelism
Goutam Sahoo
Re: Oracle enterprise stage- Degree of parallelism
goutam wrote:Hi folks,
I have a doubt on oracle enterprise stage's degree of parallelism when it acts as source.
When read method= Table (stage automatically reads in parallel).
what is the degree of parallelism here if my table is partitioned?
what is degree of parallelism here if my table is not partitioned?
when read method = user defined SQL or auto generated SQL
and and partition table property is not set stage reads sequentially. Is it sequentially per partition or runs on only one Node i.e conductor node?
when read method = user defined SQL or auto generated SQL,
and partition table property is set
I think degree of parallelism is no of partitions rather than no of nodes in ATP_CONFIG_FILE . Please confirm this.
Any idea about the above topic??
Goutam Sahoo
-
ray.wurlod
- Participant
- Posts: 54607
- Joined: Wed Oct 23, 2002 10:52 pm
- Location: Sydney, Australia
- Contact:
Learn patience.
DSXchange is an all-volunteer site, where people post as and when they can, and usually don't if they have nothing to contribute. "Pushing", as you have done, may be regarded as offensive.
If you need the answer urgently, sign up with your support provider for premium service, and learn how expensive "urgent" can be.
What research have YOU conducted in an attempt to answer your question? What results have you seen?
DSXchange is an all-volunteer site, where people post as and when they can, and usually don't if they have nothing to contribute. "Pushing", as you have done, may be regarded as offensive.
If you need the answer urgently, sign up with your support provider for premium service, and learn how expensive "urgent" can be.
What research have YOU conducted in an attempt to answer your question? What results have you seen?
IBM Software Services Group
Any contribution to this forum is my own opinion and does not necessarily reflect any position that IBM may hold.
Any contribution to this forum is my own opinion and does not necessarily reflect any position that IBM may hold.
Coul n't do a research as i don't have select privillege on system table to read a particular table thru oracle enterprise satge. What i can do is to read the parallel job developer guide and get the things clear.ray.wurlod wrote:Learn patience.
DSXchange is an all-volunteer site, where people post as and when they can, and usually don't if they have nothing to contribute. "Pushing", as you have done, may be regarded as offensive.
If you need the answer urgently, sign up with your support provider for premium service, and learn how expensive "urgent" can be.
What research have YOU conducted in an attempt to answer your question? What results have you seen?
Could you please tell me what exactly we put in partition table property to read the table in parallel.
Can a table be read in parallel even if a table is not physically partitioned?
Goutam Sahoo
Are you able to run PX jobs against any of your Oracle tables?
I suggest you try running a simple job that extracts from a single table, passes data through a parallel stage e.g. a transformer and writes it to a dataset. Observe how the job behaves.
Try running the job on one node then on multiple nodes.
Once you have understood this try the same on a partitioned table with and without the partition table option (which should be set to the table name).
When it comes to reading an Oracle table in parallel you need to define exactly what you mean. I suspect you mean can DataStage have multiple readers, but a table can be read by Oracle using parallel query even though the data is sent in a single stream to the client (DataStage, Toad or some other).
I suggest you try running a simple job that extracts from a single table, passes data through a parallel stage e.g. a transformer and writes it to a dataset. Observe how the job behaves.
Try running the job on one node then on multiple nodes.
Once you have understood this try the same on a partitioned table with and without the partition table option (which should be set to the table name).
When it comes to reading an Oracle table in parallel you need to define exactly what you mean. I suspect you mean can DataStage have multiple readers, but a table can be read by Oracle using parallel query even though the data is sent in a single stream to the client (DataStage, Toad or some other).
-
keshav0307
- Premium Member
- Posts: 783
- Joined: Mon Jan 16, 2006 10:17 pm
- Location: Sydney, Australia
-
Sreenivasulu
- Premium Member
- Posts: 892
- Joined: Thu Oct 16, 2003 5:18 am