Cheat Sheet: Control and Orchestration with KNIME Analytics Platform

Flow Variables allow for the Widget and Configuration nodes create one or more new A loop is a sequence of operations that is repeated until a condition A switch construct allows
parameterization of a workflow. A Flow flow variables & make them available at the output port. Widget is met. It has a start, an end, & a loop body of operations. A loop is you to conditionally execute
Variable is a parameter that can assume nodes create a UI item for the composite view or the KNIME implemented via a Loop Start node, a Loop End node, & a number of different sequences of
different values at different execution points in WebPortal (textbox, radio button, etc.) to create and control the flow nodes in between for the body of operations. Different Loop Start operations via nodes
the workflow & overwrite configuration variable. Configuration nodes create a UI item in the configuration nodes provide alternative ways to iterate on the input data. Different located on different
settings in upcoming nodes. dialog of a component & are not visible in the composite view or the Loop End nodes provide alternative ways to collect results. The end workflow branches. All start
KNIME WebPortal. condition can be defined in either the Loop Start node or the Loop End with a Switch Start node
Hidden Flow Variable Ports node, depending on the kind of loop. Some nodes to start & end a loop and optionally end with a
Column Selection
Each node has two hidden Flow Variable ports to accept Widget Column Selection Widget: Creates a list of selectable work on data & others on Flow Variables. These nodes can be paired Switch End node. In
incoming Flow Variables & to propagate them to the columns from the input data table in the form of a up freely - loops can start with data & end up with Flow Variables & between, a number of
upcoming nodes. To make these ports visible, right-click menu or radio buttons. The node produces the name of vice versa. parallel branches implement
the node & select Show Flow Variable Ports. Only ports of the selected column in a flow variable at its output port. various operation
Table Row to Loop End
the same type can be connected. Variable Loop Start Table Row to Variable Loop Start: Starts a loop iterating sequences.
over input data rows one at a time i.e. each iteration is Loop End: Ends a loop by concatenating the resulting
Creating a Flow Variable A switch construct works on data, flow variables, database queries, or any
Interactive Range
Interactive Range Slider Filter Widget: Creates a slider dedicated to just one row. At each iteration, the row values rows from each iteration.
1. Use a Configuration or a Widget node to create a Flow Variable at any Slider Filter Widget other port type defined via the dynamic ports of the start and end nodes.
to filter data to only include rows with values in the are converted into Flow Variables & named after the
point in your workflow. You can pair up different port types.
selected column within the specified range. The slider column headers.
1. Use any of the nodes converting data into Flow Variables.
can interact with views from other JavaScript based
3. Via the node configuration window in the Flow Variables tab, fill in a Loop End (Column CASE Switch Start: Selectively activates only one of its three
nodes in the same composite view. Group Loop Start
Group Loop Start: Starts a loop iterating over groups of CASE Switch Start
blank box with the name of the Flow Variable. Append)
Loop End (Column Append): Ends a loop by joining output ports, enabling three alternative paths for the input.
input data rows i.e. each iteration works on a different
together the resulting columns from each iteration on The active output port can be configured manually or
group of the input data. Groups are extracted from
Text Output Widget
the column containing the row IDs. automatically via the value of a Flow Variable. The IF Switch
values in selected columns, as in the configuration
String Manipulation String Manipulation (Variable): This node is the Flow Text Output Widget: Creates a paragraph of either free, node performs the same task but with only two alternative
(Variable) window of the GroupBy node.
Variables version of the String Manipulation node. preformatted, or HTML text. output ports (both can be active at the same time).
Similarly, other nodes have their own version for Flow Column List Variable Loop End Variable Loop End: Loop End nodes do not only work CASE Switch End
Loop Start
Variables like the Rule Engine Variable node & the Math Column List Loop Start: Starts a loop iterating over a on data. At the end of a loop you might want to pass CASE Switch End: Collects the results from the active one
Formula (Variable) node. selected list of columns. At each iteration, the current the results as a Flow Variable - like in the Variable among the branches connected to its input ports. The End IF
Credentials Widget column & the remaining columns are passed into the Loop End node. node works similarly & is paired with the IF Switch node.
Merge Variables: Combines Flow Variables from two or Credentials Widget: Creates fields to enter credentials
Merge Variables loop body.
more separate branches. To add a branch click the three (username & password) in the form of text boxes. The
dots in the bottom left corner. If Flow Variables with the text in the password box is masked. The node produces Variable Condition

these credentials in a flow variable at its output port. Generic Loop Start Loop End
Variable Condition Loop End: Ends a loop when a Try (Data Ports): Starts a try-catch construct to enable an
same name are collected, the Flow Variable in the top Try (Data Ports)
alternative path for the data flow in case of failure in the main
most connection is retained. Generic Loop Start: Starts a loop. It should be paired condition is met, i.e. a specific value in a Flow Variable.
with a loop end node defining the end condition. branch. One branch is defined as the main branch while the
Single Selection Single Selection Configuration: Creates a list of options other is set as the secondary branch. If execution fails in the
Inject Variables Configuration of type String in the form of a menu or radio buttons. main branch, the secondary branch is activated. It must be
(Data) Inject Variables (Data): Adds (injects) the Flow Variables
These options can be defined in the configuration dialog Parallel Chunk End closed by a Catch node.
at its Flow Variable input port into the data table at the Parallel Chunk Start: Starts a parallel loop iterating over
together with the selected default value. The node
top input port. The input data table is forwarded Parallel
input data rows in chunks of equal size which get Parallel Chunk End: Ends a parallel loop by
produces the value of the selected option in a flow Chunk Start Catch Errors
(unaltered) in the node output port. processed in parallel in the loop body. The number of concatenating the resulting data rows from each chunk. (Data Ports) Catch Errors (Data Ports): Closes a try-catch construct
variable at its output port.
parallel iterations, (chunks), is defined either automatically started with a Try node & collects the results from the active
Extract Variables
Booelan
Configuration Boolean Configuration: Creates a boolean selection for based on the size of the input data, or manually. branch.
(Data)
Extract Variables (Data): Extracts the Flow Variables an enabled/disabled flag (1/0) in the form of a Recursive Loop End: Ends a recursive loop. The top input
checkbox. The node produces the value of the selected Recursive Loop Start: Starts a recursive loop iterating on Recursive Loop End
port collects the resulting data rows. The lower input port
coming in through the input data port & produces them as Recursive the updated input table. The input data table for each
standalone Flow Variables at the output port. option in a flow variable at its output port (1 if the Loop Start collects the updated data table to be passed back to the Empty Table Switch: Provides an alternative path for the data
checkbox is enabled, 0 if disabled). iteration is the output data table from the previous iteration. Recursive Loop Start node. Also defines the loop end
Empty Table Switch
flow in case the main branch has no data rows. It activates
The first iteration works on the data table provided at the condition, such as max. number of iterations, min. the top output port & deactivates the bottom output port if
input port of the node. This is the only loop where the number of rows, & a specific value for a Flow Variable. the input table has at least one data row. It deactivates the
A Metanode or Component is a node that Metanodes just collect nodes inside and are an efficient updated data table feeds the next iteration. top output port & activates the bottom output port if the
contains other nodes. way to clean up your workflow. Feature Selection Loop End: Ends a feature selection
Feature Selection Loop Start (1:1): Starts a feature input table is empty.
Components encapsulate & abstract functionality, can have their own Feature Selection Feature Selection
loop. It sets the variable with the metric to be
Creating a Metanode or Component Loop Start (1:1) selection loop iterating on a set of input columns Loop End
dialog and can have their own sophisticated, interactive views. They maximized/minimized for best feature selection. The top Active Branch
Select all relevant nodes, right-click and select Create (features) to extract the subset that optimizes a given Inverter Active Branch Inverter: Changes the activity status of the
can be reused in your own workflows but also shared with others: via output port of the node produces the evaluation metrics
Metanode… for a Metanode or Create Component… for a Flow Variable. Used for feature selection against a given branch. If the input port is active, the output port becomes
KNIME Server or the KNIME Hub. They can also represent web pages for all column subsets. The bottom output port produces
Component. Right-clicking a Metanode or Component opens model & against a given performance metric. Together inactive & vice versa. It’s often used to force a branch to
in a Data App deployed to others via KNIME Server. Flow Variables a summary of the different feature sets with the
the context menu with a number of options such as open, with the Feature Selection Loop End node, it’s used in the produce an output even if it’s inactive & vice versa (to
cannot enter or exit a Component, unless explicity configured in the associated scores.
expand, setup, or reconfigure, and save as template. Backward/Forward Feature Selection metanode. deactivate a branch even if it’s active).
component’s input and output nodes.

Node ports
Different types of data pass through different node ports. Only ports of the
same type can be connected. Here are some examples of ports for E-Books: KNIME Advanced Luck covers these
Container
Container Input (Table): Receives a data table from the caller Timer Info
Timer Info: Reports the number of executions &
Input (Table)
frequently used data types. advanced features & more. Practicing Data
workflow. If no input is provided, the template default data table is execution times for each node in a workflow. Both
used. Similar nodes are available to exchange Flow Variables & Science is a collection of data science case
single node & total workflow execution time are studies from past projects. Both available at
credentials. The corresponding “Container Output (Table)” node reported. Execution times for nodes inside metanodes
returns the results as a data table. can also be reported.
knime.com/knimepress
Send Email
Call Remote Workflow Call Remote Workflow (Row Based): Triggers the execution KNIME Blog: Engaging topics, challenges,
(Row Based) Send Email: Sends HTML or text formatted emails
of a remote external workflow via REST. Remote means industry news, & knowledge nuggets at
using an external SMTP to a recipient - including the
stored on a KNIME Server accessed using the server URL & knime.com/blog
message & possible attachments.
credentials. Data exchange with the triggered workflow
happens via JSON format.
E-Learning Course: Take our free online self-paced
Call Workflow Service Call Workflow Service: Triggers the execution of a Container
Container Input (JSON): Receives a JSON data structure from the
Save Workflow
coursesto learn about the different steps in a data
workflow stored in the LOCAL workspace or on KNIME Input (JSON)

Server. Data exchange with the triggered workflow can caller workflow. If no input is provided, the template default JSON Save Workflow: Saves the (also partially - up to here) science project (with exercises & solutions to
happen via any port type via the Workflow Service structure is used. The corresponding “Container Output (JSON)” node executed workflow. test your knowledge) at
Input/Output nodes. returns the results as a JSON structure. www.knime.com/knime-self-paced-courses
Workflow Service Workflow Service Input: Receives data via any port type
Input from the caller workflow via the Call Workflow Service GET Request GET Request: Calls a REST service in GET mode. The node can send Create Folder KNIME Hub: Browse and share workflows, nodes,
node. Enables an efficient data exchange between one single service request set in the configuration window, or multiple Create Folder: Creates a new folder and outputs the and components. Add ratings, or comments to
workflows only, excluding third-party software. The service requests stored in a column of the input table. Responses are folder location as a flow variable of type Path. other workflows at hub.knime.com
corresponding “Workflow Service Output” node returns the saved in the output data table. Options to set authentication, request
results to the caller workflow. header, & response header are available.
KNIME Forum: Join our global community &
Call Workflow
Call Workflow (Table Based): Triggers the execution of a KNIME Server
Create File/Folder
Variables
engage in conversations at forum.knime.com
(Table Based) Connector KNIME Server Connector: Connects to a KNIME Server using the Create File/Folder Variables: Creates a list of Path
workflow stored in the LOCAL workspace or on KNIME
server URL & credentials. After the connection has been created, type Flow Variables pointing to files/folders relative
Server. Data exchange with the triggered workflow can KNIME Server: For team-based collaboration,
new directories on the server can be created & remote files can be to a selected base location.
happen via data tables, Flow Variables, or credentials via
accessed, created, & deleted. automation, management, & deployment check
the respective Container Input/Output nodes. out KNIME Server at www.knime.com/knime-server
KNIME Press
Extend your KNIME knowledge with our collection of books from KNIME Press. For beginner and advanced users, through to those interested in specialty topics such as topic detection, data blending, and classic
solutions to common use cases using KNIME Analytics Platform - there’s something for everyone. Available for download at www.knime.com/knimepress.

KNIME ®
BEGINNER·S
LUCK

Decision
Tree Learner
File Reader Partitioning

Decision Tree
training to Predictor Scorer
original 80 vs. 20 predict income
data set

attach class confusion matrix

probabilities + scores

A Guide to KNIME Analytics Platform for Beginners

Authors: Satoru Hayasaka and Rosaria Silipo

KNIME for Life Sciences

A Collection of Use Cases

Blend & Transform Validate & Deploy

SECOND EDITION
Data Blending with KNIME
Model & Production Consume &
Visualize Creation Production Interact
Process

Optimize & Capture Monitor & Update

KNIME®
Rosaria Silipo & Lada Rudnitckaia

© 2022 KNIME AG. All rights reserved. The KNIME® trademark and logo and OPEN FOR INNOVATION® trademark are used by KNIME AG under license from KNIME GmbH, and are registered in the United States. KNIME® is also registered in Germany.