Indeed, there are a few news in the pages of the book! Let’s discover them together.
- Chapter 10 has been dropped. The memory handling part has been absorbed into Chapter 1 and the batch execution has been dropped completely.
- Chapter 1. The installation of the R software has changed a bit to be able to work with the KNIME Analytics Platform. Such changes have been described in section 1.4.
- Chapter 2. The database connections and operations have changed greatly in the last versions of the KNIME Analytics Platform. Basically, you can now connect and extract data from a database with different shades of modularity.
- The first original monolithic approach includes nodes like the “Database Reader” node. These nodes can define connection details and data extraction SQL queries all in their configuration windows. These nodes also transfer the resulting data back into the KNIME workflow.
- Thinning down the monolithic approach, the “Database Table Connector” node still connects to the database and builds a SQL query for the data extraction. However, it does not run the SQL query and does not import the results into the KNIME workflow. In order to do that, a “Database Connection Reader” node has to be connected to follow the “Database Table Connector” node. This approach leaves the SQL query output port of the “Database Table Connector” node open for a number of in-database processing nodes, which simplify the building of an appropriate SQL query through their graphical user interface.
- But the most modular approach of all includes a general “Database Connector” node (or alternatively a dedicated connector node), followed by a “Database Table Selector” node, a number of in-database processing nodes, and finally a “Database Connection Reader” node, to import the final results into the KNIME workflow. The additional advantage of this very modular approach is in the flexibility of the database of choice. If I change the database to connect to, I just need to change the initial connector node. Everything else that follows would still work.
- Chapter 4. Section 4.5 reports the differences between a metanode and a wrapped node. It also introduces the “Layout Wrapped Node Wizard” that helps with the relative positioning of Quickform GUI objects on the KNIME WebPortal.
- Chapter 5. The new R nodes have been added here: “R Source” and “R Metanodes”
- Chapter 7. Loop sections have been rearranged based on the frequency pattern of usage; a better example has been used to describe the loop till a condition at the end is verified (section 7.8); and the recursive loop has been newly introduced in section 7.9.
- Chapter 8. Now chapter 8 also mentions CASE Switch blocks for variables and models besides data tables.
The book is still available and can still be purchased through the KNIME Press web page https://www.knime.org/knimepress/the-knime-cookbook