Is there a way to save a template table in the new Workflow service input node as in the Container input (table) node?
As far as I can see in KNIME 4.5 the Capture workflow and Workflow writer nodes generate a new workflow with Workflow service input nodes and won’t use the Container nodes anymore.
yes, that’s not possible with the Workflow Service Input node. And since I wasn’t involved in the implementation of these nodes I can’t tell at the moment whether it’s been a conscious decision. But I will check with the developer.
For the time being you can still add ‘template’ data by just connecting a node to the optional input port of the Workflow Service Input node. To make it work using Integrated Deployment (i.e. Capture and write workflows) you will need to explicitly add the Workflow Service Input node like this:
Alternatively, you can still revert back to using ‘Container Input (Table)’ nodes and the Workflow Writer (and Deploy to Server node) still allows you to select those:
However, if the written workflow is exclusively used to be called from other workflows I would not
recommend to use the Container I/O (Table) nodes mainly because the data transfer is less efficient.
Thanks for your detailed answer. I have started adapting the example so it looks more like what I had in mind. And the Capture nodes can be used to hide some complexity but the output workflow has still sample data to be executed.
I would also need to store some files in the data area (see screenshot below). But these are deleted in the workflow created by the Workflow writer. Is there a way to keep these files?
// Alternatively, you can still revert back to using ‘Container Input (Table)’ nodes and the Workflow Writer (and Deploy to Server node) still allows you to select those
Thanks for pointing me to this tab: I missed it.
The data area you are writing the ‘alternate input data’ to belongs to the actual workflow project (i.e. the workflow that hosts the Workflow Capture Start and -End nodes). And yes, that’s right, the data of the ‘workflow project data area’ is not copied into the data area of the newly written workflow. Though it’s an interesting idea. However, in your example the captured workflow segment doesn’t seem to make use of the ‘alternate input data’ from the data area anyway, or does it? I.e. what’s the actual use case?
The only way to store static data with the captured workflow is to use this pattern:
(as probably already noticed the ‘Table Creator’ node is replaced by a ‘… Reference Reader’ node when the captured workflow segment is written via the Workflow Writer).
Is there a use case where this approach doesn’t help?