Replacing the value of a column by another value of the same column

[rsison]

Hi I’m new in Knime and want to know how can I replace the score value of partner by the score value of customer focused. Please see image below. Thank you!

[AlexanderFillbrunn]

Hi,
That is probably possible, but how do you know which “customer focused” line belongs to this partner line? Is it always 2 rows ahead?
Kind regards,
Alexander

[iCFO]

While this is definitely possible via the Lag node or a multi row formula in Column Expressions node… It looks to me like the table would be cleaner if it was pivoted or joined differently. It would be much easier to work with overall if the underlying structure placed them on the same row.

[rsison]

Hi Alexanxder,

We can use the stat column to capture the correct row.

Thanks!

[rsison]

Yes I agree but the one who provides the data can’t do it in R.

I was looking at several possible nodes like rule engine, constant var column, and string manipulation but I’m not sure how to execute it

[AlexanderFillbrunn]

Hi,
I think you won’t need a Lag Column. Just add a Row Filter that keeps only rows where attribute=“Customer focused” and another one where attribute=“Partner”. Then Use a Joiner to join the two filtered tables by the column “stat”. You might want to use the “Left Join” mode to keep all Partner entries and not just those with a matching “Customer focused” entry. The joiner’s config window also has a second tab where you can select the columns to keep from both tables. Maybe you will need a Rule Engine to coalesce the two score columns you will have (partner and optional Customer focused), but that is basically it. In the end you can concatenate the treated rows back with the “Helps achieve my goals” entries.
Kind regards,
Alexander

[iCFO]

You can always transform the structure that he can provide you via R into a clean table using KNIME tools. If you don’t need a cleaner table for further downstream analysis, then the method that @AlexanderFillbrunn provided will get you there. If you prefer to just transform the table to an easier structure, then you can post a workflow or excel file that has your columns and we can help on that end.