I have a sheet read from excel file with series of column header(more than 100) like “2023-06-13”, “2023-06-14” and need to output these header as the format of “06/13/23”.
I tried to transpose it to a column to change the format, but stuck when “string to date&time” required a column name, which don’t exist after transpose.
What @ipazin has provided is clearly the solution to your specific problem and as you can see, Column Rename (Regex) is a really useful node for these situations. Regex has many uses in KNIME and elsewhere, and I can recommend the regex101.com site for trying out and exploring regex.
As you mentioned transposing tables to modify the column headers, I thought I’d also add some further information on how you can modify the column names using table transposition and various transformation nodes, which could be adapted for situations where regex rename alone would not be sufficient.
btw the “Transpose” node in KNIME 4.x becomes the “Table Transposer” node in KNIME 5.1
NB. Please don’t additionally try to mark this response as the solution though, even if it works as only one answer can be a solution on a thread, and I think that @ipazin’s answer is the best solution for your question.
@takbb thanks for the update.
I love KNIME and I really appreciate the great work KNIME team to be honest I do not like the renaming of etablished Nodes in the new version. This leads to confusion. I think you guys should switch back to the old names in 5.2
br
I was thinking something similar, although I can understand the push towards nodes having more meaningful names. But yes, when you just “know” which node you want and can no longer find it it gets frustrating.
I’ve been looking around at least for a list of node changes. The release note for 5.1 shows the new nodes, but is there any documentation on node renaming? I might have missed it. Maybe one for @ScottF
Scroll down to “Nodes changing in KNIME Analytics Platform 5.1.0”
and the sub-section “Nodes renamed”
I note that searching on the hub for “Insert Column Header” now shows it as the new name but there is nothing there to show it had an old name. That’s a little confusing for us “old timers”. and tbh, it’s going to be very confusing for “new timers” when they read solutions on the forum and then cannot work out what nodes we’re talking about. It would be nice if somehow the hub could note name changes, but maybe that’s a stretch technically.
I think it would also have been useful in 5.1 that if when we try to search for a node using an old name, that it at least presents us with the new node. That might then give us a clue. When I finally make the full move to 5.x.x, I think I may have to still keep a copy of 4.7 available so that when I can’t find a node, I’ll drop it on a workflow in 4.7, export it into 5.x.x and see what it becomes!
@Daniel_Weikert and @takbb I wanted to add something about this part of the discussion, since it has come up internally as well.
If you need to find deprecated nodes in the UI, there is a bit of a hidden feature: using //deprecated in the node search will make those nodes visible:
Caveat: this feature that might disappear in the future since it’s undocumented and tucked away. But maybe for now you find it useful when these scenarios come up.
Thanks @ScottF , that will be useful for deprecated nodes.
If there could be a similar hidden feature that can find a node based on the old name such as searching for “Insert Column Header” and having it show up “Column Renamer (Dictionary)”, that would be great too or maybe this could be a job for K-AI… It had mixed results at telling me new names for nodes, but maybe in future it could learn the history of each node?
I think as users, they may not realized it’s the version change that preventing them from finding the previous node. If the result could show the new node with a note of old name, when searching with old name, the experience might be more seamless.