Updates involve not only compliance with the latest KNIME version, but a few more structural and analytics related changes. This post lists and describes the most important of them.
Compliance with KNIME Analytics Platform 3.0
All workflows, pictures, screenshots, node descriptions, and other instructions have been updated to be compliant with the new functionalities and look of the KNIME Analytics Platform 3.0. This paragraph is just a few words, but it describes a lot of new things!
Data and Workflow Structure of the Download Zone File
(section 1.16, page 35)
The Download Zone is a file including all workflows and data used in the book, both for demos and for exercises. It is a zip file to be downloaded at purchase time (link to the zip file is included in the email with the book delivery). With the latest new versions of KNIME, it is possible to import data and workflows into the KNIME Explorer panel. Accordingly, the newest Download Zone file includes workflows and data that can be both imported directly into KNIME, without any additional step.
- Download the Download Zone zip file onto your machine (do not unzip it!)
- Open KNIME Analytics Platform
- In the Top Menu, select “File” -> “Import KNIME Workflow ...”
- In the “Import Workflow” window, select option “Archive File” and open the Download Zone zip file
- Select destination workflow group in the KNIME Explorer panel
- Make sure that all workflows and data are selected in the “Import Elements” preview
- Click “OK”
The knime:// Protocol (page 54)
Still exploiting the new data options in the KNIME Explorer panel, we introduced the knime:// protocol in all example and exercise workflows. The knime:// protocol allows to use the current workflow folder as a URL reference. So, URL “knime://knime.workflow/../../KBLdata/<filename> points to a file located two levels up from the workflow folder and under folder KBLdata. More details on the knime:// protocol are available at page 54 of the book.
The new Column Selection Panel (page 59)
At page 59, the description of the “Column Filter” node allows for the introduction of the new column selection panel, used in all nodes where column selection is required. Indeed, it is now possible to select columns manually (as it was before), by type, and by regex or wildcard based expression. These last two options are very useful when the high number of columns makes the manual column selection too time consuming.
Database Nodes (section 3.7, page 91)
Another big change in the KNIME Analytics Platform 3.0, which has been reported in the book, is about the database nodes.
There are two ways now to access a database with KNIME (section 3.7, page 91):
- Using a standalone “Database Reader” node (or a “Database Writer” node)
- Using a database connector node (generic or dedicated) followed by other nodes for in-database processing and to extract data into the KNIME platform.
All sections dealing with database nodes have been changed to reflect one of the two approaches, most frequently the second one.
In section 3.8 about nodes for data visualization, I have reported the three options available at the moment with KNIME:
- The old option with Data Views nodes (now called just Views nodes)
- The JFreeChart nodes, which are just a copy of the Views nodes with no interactivity and the production of a static image to be used for reporting
PMML Introduction (section 4.3, page 131)
The PMML standard is introduced at page 131 already. Indeed, it is possible to export data transformations as well as predictive models into PMML format and combine them together. So, before I dive into the “Missing Value” node, I give a short introduction of the PMML format.
GroupBy Aggregation Tabs (section 5.2, page 175)
The description of the “GroupBy” node has been extended with the additional new aggregation tabs. Complying with the new column selection panel, there are three options now to select the columns for value aggregation in the “GroupBy” node:
- Manual selection
- Type based selection
- Pattern based aggregation (based on a Regex or wildcard expression)
The new Context Menu for Metanodes (page 196)
Options for Metanodes in the context menu have been grouped together in a context sub-menu explicitely dedicated to metanodes. This is described at page 196.
Updates to comply with BIRT 4.0 (Chapter 6)
The whole chapter 6 has been reviewed to comply with the latest BIRT release 4.0. Notably, the main change is in the absence of the Preview tab in the BIRT environment.