This scenario describes a two-component Job in which the rows of an ARFF file are read, the delimited data is selected and the output is displayed in the Run view.
An ARFF file looks like the following:
It is generally made of two parts. The first part describes the data structure, that
is to say the rows which begin by
@attribute and the second part comprises
the raw data, which follows the expression
Drop the tFileInputARFF component from the Palette onto the workspace.
In the same way, drop the tLogRow component.
Right-click the tFileInputARFF and select Row > Main in the menu. Then, drag the link to the tLogRow, and click it. The link is created and appears.
Double-click the tFileInputARFF.
In the Component view, in the File Name field, browse your directory in order to select your .arff file.
In the Schema field, select Built-In.
Click the [...] button next to Edit schema to add column descriptions corresponding to the file to be read.
Click on the button as many times as required to create the number of columns required, according to the source file. Name the columns as follows.
For every column, the Nullable check box is selected by default. Leave the check boxes selected, for all of the columns.
In the workspace, double-click the tLogRow to display its Component view.
Click the [...] button next to Edit schema to check that the schema has been propagated. If not, click the Sync columns button.