Arbitrary numbers of flow variable arguments to various nodes

[ajackson]

There’s an issue with setting flow variables for many nodes where there can be an arbitrary number of arguments, that is platform-wide.I reported on the Column Resorter about 18 months ago.

KNIME AP 5.3 - Sorter node - Missing functionality with Flow Variables

With KNIME AP 5.3 I noticed the Sorter node’s Flow Variables have changed, removing some functionality. Is there a simple viable alternative, or can this prior functionality be restored? Prior to 5.3 I was able to set Flow Variables to get a string list/array, like so: [image] This would then yield: [image] With 5.3, the Sorter node’s Flow Variables look like this: [image] Which yields: [image] This feels like a step backwards. To get flow variables out of a Sorter node, my only opt…

Other nodes have had the problem for many years:

While some nodes are absent pretty much all FV config:

As someone who relies on flow vars on a daily basis, the ability for nodes to accommodate them makes a huge difference in what can be down (and how).

Is this going to be addressed as the new UI stuff matures? …or is this the intention going forwards (or rather backwards, as it seems!)

Cheers, AJ

[fe145f9fb2a1f6b]

having the option to simply pass arrays for column selection, or “blocks” of operations e.g. in Row Filters has been long gone…

instead they imposed a stupid system in which you need to know either ahead of designing the workflow, how many columns you want to manipulate, or you need to wrap everything into a loop.

in many cases these only show up dynamically once you configured the node and then are hardstuck as shown below.

their idea of assigning flow variables to a tree like model:

var_whatever
1:
 - var1_ColumnName
 - var1_ColumnOperation
 - var1_ColumnFilterValue
2:
 - var2_ColumnName
 - var2_ColumnOperation
 - var2_ColumnFilterValue
3:

instead of simply setting 3 arrays [ColumnNames], [ColumnOperations], [ColumnFilterValues]. this is also in conjunction with their modern ui.

Arrays and Lists are neglected datatypes for Knime anyway - there are about 3 nodes working with them (subset nodes) - and fewer nodes supporting them via flow variables since webui port.

it feels like the people at Knime that actually used it for data related work and automation, left the company 4-5 years ago…

[ajackson]

It seems like a JSON or format could overcome these issues quite naturally.

[fe145f9fb2a1f6b]

arrays are present, have been used in the past, and work perfectly fine for this case. further they are datatype safe (which json isnt) and a lot easier to handle within Knime compared to JSON

[ajackson]

I don’t think there’s a need to worry about datatypes with Flow variable configuration. All that’s happening in the Flow Variable dialog is choosing a flow variable to use as a setting for a node.

[fe145f9fb2a1f6b]

until the node expects a path flow variable but you select a string, or expects a integer but you select a double…