Calls another workflow and obtains the results for further processing in this workflow. The workflow can receive inputs via Workflow Input nodes and return outputs using Workflow Output nodes.
Each Workflow Input node in the workflow to be called will create an input port on this node when finishing the configuration of the Call Workflow Service node. Similarly, each Workflow Output node will create an output port on the node.
The called workflow can be local or remote. If the workflow is remote, e.g., on the KNIME Hub, the execution will also be remote. When the workflow is local, the execution will be performed on the machine on which the Analytics Platform is running. In contrast to the Workflow Executor node, this node does not require the workflow to be read using a Workflow Reader node.
The difference between this node and the Call Workflow (Table Based) node is the set of supported workflow input and output nodes. This node supports the Workflow Input and Workflow Output nodes, which support arbitrary port types and are more efficient than the various Container Input and Container Output nodes. The container input and output nodes on the other hand expect and produce data in a format that can easily be produced by third-party software, whereas Workflow Input and Workflow Output nodes are designed exclusively to be used by other KNIME workflows.
To define which Workflow Input node receives data from which of this node's input ports, each input node defines a parameter identifier. The parameter identifier is supposed to be unique, but it might happen that a workflow has multiple input nodes defining the same parameter name. In this case, KNIME will make the parameter names unique by appending the input node's node ID to the parameter name, e.g., "input-table" becomes "input-table-7".