With Load and Unload Utilities - Access Module

Teradata Tools and Utilities Access Modules User Guide

Product
Access Module
Release Number
15.10
Language
English (United States)
Last Update
2018-06-06
Product Category
Teradata Tools and Utilities

With Load and Unload Utilities

Figure 4 shows how the Teradata Named Pipes Access Module works with client load and unload utilities.

Figure 4: Data Flow Between Named Pipes and Load/Unload Utilities

A writer process writes an input data stream to a specified pipe either directly or through the Data Connector API. In response to requests from a reader process, the Teradata Named Pipes Access Module does the following:

  • Reads the data from the specified pipe
  • Copies it to the fallback data file
  • Submits it to the reader process
  • The Teradata Named Pipes Access Module can also read data from an ordinary file or from a file descriptor file system device on UNIX systems. However, using the module with an ordinary file is not recommended and does not yield the same performance as using a flat file without the access module.

    The Named Pipes Access Module supports only read operations from a named pipe. You cannot use the Named Pipes Access Module to write to a pipe. Additionally, the access module is a block data transfer module; it is neither data‑type nor record sensitive.

    Writer Process

    The writer process can be a client extraction utility, such as FastExport, or any other application, data source, or device that can provide data in a format supported by the reader process. On UNIX systems the writer process must run on the same system as the reader process. UNIX pipes cannot span a network.

    The interface between the writer process and the Named Pipes Access Module can be either a named pipe or the Data Connector API. On SPARC systems running Oracle Solaris, the interface can also be unnamed pipes. However, the Teradata Named Pipes Access Module does not support interconnection through unnamed pipes on the following operating systems because these systems do not have file descriptor file systems:

  • HP‑UX
  • IBM AIX
  • Linux
  • Windows
  • Reader Process

    The reader process can be a client load utility, such as Teradata FastLoad, Teradata MultiLoad, or Teradata TPump. Because Windows named pipes can span networks, the reader and writer processes can reside on different network‑connected Windows systems.

    The Data Connector API provides the interface between the reader process and the Teradata Named Pipes Access Module.