Hi,
I can’t help myself with the following problem in KNIMA.
I have a table which contains for example 5 columns.
One column is ID and the remaining 4 are values.
If the table will have 4 rows so i need to multiply the number of columns times 4.
Because duplicates would arise so I need each column to have a suffix where it would be clear in which order the column is.
See the example in the attachment.
It’s NOT Unipovitng nor Pivoting in my opinion.
It should be done in a loop but I don’t know how.
Of course there are millions of records in the table .
The ID is repeated 4 times each time.
Hello,
I’ve worked on it a bit and you’re right , it just needed a bit of work.
The bottom line was that I need to use two different cycles in my opinion and via rowID create a new identifier which will just represent the newly created record.
Because it was 7 million rows, I had to sort it before the transfomation so it wouldn’t affect the performance
If you use the GroupBy and Cell Splitter nodes, you won’t need an extra loop.
You can use the ID as the grouping column and concatenate the values in the selected coulmns and then use a Cell Splitter in a loop to split the values in columns.
Hello,
I didn’t quite understand how your design works.
I finally took the plunge and after a few tries I was able to get it running.
GroupBy has to be set up correctly because “set” and “ist” don’t work the same.
Ok cell spliter doesn’t work in this case but Split Collection column works beautifully . Splitting Multiple Collection Columns – KNIME Community Hub
My solution did work after three hours but I still had to redo the object types because it all turned into text.
Now it took 50 minutes .