I would like to map the values into new columns with the keys as headers. I tried the ‘Cell Splitter’ already but that works best for a fixed split with a fixed output.
You can use Cell Splitter, but there are more operations to be added to accomplish what you want to do.
The main challenge here is that your records might not all have the same keys. For this type of structure, I usually use JSON, as that’s exactly how a JSON “structure” is, and also because converting JSON to a table, Knime will automatically create the necessary columns dynamically.
If I look at your data, I am assuming that each pair is delimited by a semi-colon (;), and that a pairing of key value is determined by key=value format. I am assuming it’s 2 lines, therefore 2 records, is that correct? I’m making this assumption because I see the key Parent repeating.
I have put something together based on these assumptions, and here are the results:
Hi @mpreusse , yes, it should be ok without chunking/batches or parallelization. There’s really not much happening there. They’re straight forward operations.
