Hi,
maybe I don’t understand the node correctly?
When I use “Unary Union” the results looks fine, but I loose all of the other columns (other than geometry). Is that intended by design?
Second part of the question: the result is a multipart polygon. Is there an option to create one single polygon output?
Thanks and cheers
Hello @Awiener,
Thank you for contacting the KNIME Forum.
Do I understand correctly, that your goal is to retain (some) information from the columns that are not type of “geo” in the output of the <unary union>?
The unary union does unite all objects under one multipart object in its output. As there is no option to handle the grouping information for the remaining columns you might want to process the metadata with a <groupby> node and join that to your generalized multipart “geo” object.
To have a bit more control, of what geospatial data you want to generalize, you might want to use the <dissolve> node. In here, you can define which column your generalization bases on.
E.g. in your example it could be the LAND_... column (1, 2, 3, 5, …). In that case, you would get as many singlepart or multipart objects (depending on how many of those entities are present) depending on the number of unique values of that column.
If you dissolve on the COUNT... column, you will end up with one multipart object (as long as “Germany” is the only entry).
Please have a look at this example, as this might explain this to you a bit better:
I must admit, I did not understand your second question entirely, since this is heavily dependent on the information you want to generalize.
There is a <multipart-to-singlepart> node that allows you to split the multipart objects based on a categorical column. That might help in your case. You might want to check out this example workflow to get a better understanding about it:
Best regards
Thank you, @kevin_sturm
I am trying to open your workflow, but it’s doing nothing after I load it:
When I click cancel, Knime crashes. Is there something I could do in order to work with your workflow?
I guess it has somehting to do with
However that is company regulation 
Hi @Awiener,
It might be very likely that your company regulation (making your directory read-only) does affect your observed behavior.
May I ask how you have tried to open the workflow? Did you double-click on knwf file or did you import the workflow via the interface of the KNIME Analytics Platform?
Best regards
Kevin
Hi @kevin_sturm
I’ve tried both scenarios.
If that means, that with security suggestions from your side it’s not possible to open shared workflow, there is imho a deeper issue.
I guess (and that’s just a guessing) that your workflow is trying to search for extensions I don’t have installed. And there it stucks, because as a (limited) User I am not allowed to install them.
So imho there should be a message, that the workflow cannot be installed without these extensions and that an admin must grant access to the installation process.
What do you think?
Cheers
To the Union use case:
Yeah, I don’t understand why there is a loss of attributes when the tool is used. Of course I want to work with all the attributes in the following nodes. But I got your point, that there must be a group by and join node afterwards in order to get my desired result. Thanks for this clarification. Same with dissolve which would than be the next node.
Generally speaking, I find it quite “interesting” in the latest releases, that more and more older nodes are substituted with more easier and less function-rich (aka “dumber”) nodes which can only do a single thing compared to older nodes which where much more of a workflow in itself. (Sounds like a more UNIX approach to me 
)
So that it’s necessary to use more and more nodes in order to get the same result than with older nodes which where much more function orientated. Is that just my opinion or is that a general road map?
Speaking of the Union node from above: I now need at least 3 nodes to get a result, that could be implemented in one single node with 3 tabs. 
And yes, I do understand that you incorporate Python functions in the nodes, but then a User sometime looses focus, what and when to use which node (for me, Union should at least keep all the attributes).
So documentation is getting more and more important to understand, which node to use next to achieve my desired result. However I don’t see the “In order to achieve result X, use Y as next node” part (very often) in the docs. Same with examples, which would really help to understand what a node is doing exactly.
Just my 50 Cents …
I’ve understood the concept now. Thanks for your input!
Have a great New Year 2025! And thanks a lot, you are all doing an amazing job. I love Knime 
Cheers and sorry this post got so long.