Operators and Players

Post questions here relative to DataStage Enterprise/PX Edition for such areas as Parallel job design, Parallel datasets, BuildOps, Wrappers, etc.

Moderators: chulett, rschirm, roy

Post Reply
Pravenai
Participant
Posts: 49
Joined: Mon Apr 06, 2009 1:32 am
Location: USA

Operators and Players

Post by Pravenai »

What is the difference between operators and players?
ray.wurlod
Participant
Posts: 54607
Joined: Wed Oct 23, 2002 10:52 pm
Location: Sydney, Australia
Contact:

Post by ray.wurlod »

An operator is an executable piece of code - an object of the APT_OPERATOR class. A player is the process that executes that code.
IBM Software Services Group
Any contribution to this forum is my own opinion and does not necessarily reflect any position that IBM may hold.
jhmckeever
Premium Member
Premium Member
Posts: 301
Joined: Thu Jul 14, 2005 10:27 am
Location: Melbourne, Australia
Contact:

Post by jhmckeever »

Operators are logical representations of individual executable components between which your data flows. A script which describes your job's operators and the flows between them is called the OSH script, and is what your parallel DataStage job gets compiled to.

How multiple instances of each operator are executed across a multi-processor architecture (according to your node configuration) is the job of the SCORE which describes how this is achieved using Conductor, Section Leader, and Player processes.

Players are described within the context of the score here:
viewtopic.php?t=125205&highlight=players

J.

PS. I'm sure Ray'll be along in a few minutes with a much more concise explanation! :-)
<b>John McKeever</b>
Data Migrators
<b><a href="https://www.mettleci.com">MettleCI</a> - DevOps for DataStage</b>
<a href="http://www.datamigrators.com/"><img src="https://www.datamigrators.com/assets/im ... l.png"></a>
jhmckeever
Premium Member
Premium Member
Posts: 301
Joined: Thu Jul 14, 2005 10:27 am
Location: Melbourne, Australia
Contact:

Post by jhmckeever »

Ha - He beat me to it!
<b>John McKeever</b>
Data Migrators
<b><a href="https://www.mettleci.com">MettleCI</a> - DevOps for DataStage</b>
<a href="http://www.datamigrators.com/"><img src="https://www.datamigrators.com/assets/im ... l.png"></a>
Pravenai
Participant
Posts: 49
Joined: Mon Apr 06, 2009 1:32 am
Location: USA

Post by Pravenai »

ray.wurlod wrote:An operator is an executable piece of code - an object of the APT_OPERATOR class. A player is the process that executes that code. ...
Is it true that a player is created for every stage on every node? If yes does that mean the number of operators created in a job (including the instances) will be equal to the number of players (unless no operators are combined into one)?
ray.wurlod
Participant
Posts: 54607
Joined: Wed Oct 23, 2002 10:52 pm
Location: Sydney, Australia
Contact:

Post by ray.wurlod »

It's not necessarily true. Even after you exclude operator combination, sometimes there are more operators then stages. For example, overwriting a Data Set creates two operators (called a "composite operator"), one to delete the Data Set, the other to copy data into it. A normal lookup generates a composite operator made up of LUT_CreateOp to create the lookup table in memory and LUT_ProcessOp to perform the actual lookups. Finally, when the score is composed, other operators - such as tsort and buffer operators - may be inserted while some stages - such as Copy and Modify stages - may end up generating no operator at all.
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