VM Pipelines

Overview
Based on a much-enhanced version of the Unix concept of combining a series of simple programs to solve a data processing problem, Pipelines has been hailed as the most significant enhancement to VM since the introduction of REXX. Not only is it powerful, but it typically out-performs the traditional EXECIO and XEDIT data processing techniques. This course uses a workshop approach to quickly train all apprentice "plumbers".

Course objectives
To provide a rapid introduction to the power of Pipelines, and to provide the confidence that will allow the attendee to experiment further.

By the end of this course you will have learned
How a Pipeline stage works. How to read, write, select and process data within a single Pipeline. How a multi-stream Pipeline works and how to connect the streams. How to handle Pipeline stalls. How to develop simple user-written Pipeline stages written in Rexx.

Who should attend
Any user of VM.

Pre-requisites
REXX Programming or equivalent experience.

Course synopsis

• Pipeline Concepts
• Using a series of stages to process data. How a Pipeline stage works to avoid storage-expensive data buffering, and when stages like SORT buffer it.
• Single-Stream Pipelines
• Device driver stages such as CMS, CP, to replace slow EXECIO and Rexx processing. Basic record selection using TAKE, DROP, FIND, LOCATE, ALL, PICK, etc. Using SPECS to rearrange data.
• Multi-stream Pipelines
• Design principles. Using gateway stages FANIN, FANINANY and FANOUT. Using labels to connect separate pipelines. Dealing with Pipeline stalls and out-of-sequence records.
• User Stages
• Basics of writing user stages in Rexx using PEEKTO, READTO and OUTPUT. Example stages.