EDOBE XDOM PMML Manual de usuario descargar pdf (Pagina 50)

THE TIBCO SPOTFIRE PIPELINE

Overview The TIBCO Spotfire Pipeline is a C++ system for accessing and

manipulating very large data sets.

The core of the TIBCO Spotfire Pipeline system is a set of C++

classes representing Buf, Proc, and Pipeline objects. In order to

create entirely new Proc components, it is necessary to implement

them as new C++ classes. This chapter describes the C++ classes

used by the TIBCO Spotfire Pipeline system, and explains how new

components can be implemented. New C++ Proc classes need to

follow certain rules to work within the pipeline.

The Java package

com.insightful.cnkjava contains peer classes for

each of the classes presented here. The architecture description here

is also informative for the straight Java programmer. For Java-

oriented method descriptions, see the javadoc in doc/javadoc.

Basic Objects A pipeline is composed of only a few different types of objects.

Proc

A Proc is a data processing object. There are many different types of

procs, used for performing different operations, including reading

data from a file or database, performing a transformation on some

data, accumulating count information, or constructing a linear model

from a data stream.

Buf

A Buf is a data buffer. A buf has multiple named columns which can

contain elements of different types, and N rows of data. A buf is used

like a circular buffer.

When a proc produces new data rows, they are written to a buf. One

or more procs read the data rows from the buf, in the order that they

were written. The only way that one proc in a pipeline can send data

to another one is through an intermediate buf object.

1 2 ... 45 46 47 48 49 50 51 52 53 54 55 ... 97 98

Sin comentarios

EDOBE XDOM PMML Manual de usuario Pagina 50