I’ve noticed that the Table Editor (JavaScript) node has recently been flagged as Legacy, along with the Table Viewer (JavaScript). However, while there is a new Table View node, there is no equivalent for the Table Editor. So the only way to edit a table in a data app now is to use a legacy node. This suggests that this functionality is no longer supported in KNIME.
The ability to edit tables in data apps is very important to our business and many of our clients, and the loss of this feature would require us to rethink many of our KNIME applications and either find non-KNIME solutions or develop our own table editor node (which will be time consuming), or hope that a trusted third party develops one.
Is a new Table Editor being developed, or is this no longer a core KNIME feature?
I am Femke from the UX Team at KNIME.
…as you already asking for the Table Editor node, I would like to know if you can explain what you liked in the old legacy node and what you might miss? You can describe some of your usecases / problems that this node solves as well.
I would love to hear your feedback, if you have time!
Femke
Hi @FemkeVegel, thanks for seeking my feedback. Generally the old node works pretty well. This single biggest issue is well known – namely, the ability to paste illegal values (via Ctrl-V) that otherwise could not have been entered via a domain-controlled dropdown. This completely undermines the domain constraint as a way of excluding invalid input.
On a related matter, it would be great if a column could be configured to allow free text while also providing dropdowns, instead of just one or the other. This could be combined with some kind of auto-complete, similar to Excel.
Another limitation with the current node is how it behaves with the refresh button. If there is a refresh button upstream, the table reverts to its original values every second time the button is pressed. I can understand why this happens, and I suspect that the node would need a special setting to retain the user’s edits, but if this was possible, it would provide some new options for data app design. Presently, the node’s contents must be constantly written to and read from a DB table or file to be retained across refresh events.
The ultimate dream feature would be dynamic domains across columns, so the dropdowns available in one column depend on the value entered in another column. This would reduce the need for validation rules after submission. It might be difficult to achieve, but in theory it could be done with the right input (e.g. a table listing allowed value combinations for different column pairs).
Also, the current node is not good for entering longer strings into a cell. The ability to pop open a larger text box, just like the one for reading long strings in the new KAP node monitor, could solve this problem.
I can feel all your points. I have added your feedback to our Ticket so that it stays visible. We will try to add some points to the first node release, but in general I cannot promise to include all. You know the drill, things go step by step
In general you can always @ me here, if you want to give feedback for existing Nodes.
Cheers, Femke
