Pivot help...maybe

Please see the graphic here

There are five “view_token_arr (0-4)” column headers currently in the data set.

The cell data pools into the columns simply based on the first Numeric ID found. For example column arr0 has ID 5005, 5001, 5004, all in the same column.

I would like to make the “numeric IDs” currently appearing as cell values for the various view columns into column headers (ie: new column headers of “5001”, “5004”, “5005”, etc.), keep the “ListingId” column just as it is helping to identify each property as a row, and place “PSF above ground” as each cell value.

This way each view column scores each property when the column header is only one of several IDs instead of the mixed values each column currently has. I have just bombed out trying to get this to work.

Once this works I need to do it for many other columns using the functioning strategy.

Here’s the data set CompleteQuantitativeVariables.xlsx (658.8 KB)

thank you

@creedsmith you might have a look at this approach. Combining One to Many and Math Formula over multiple Rows to transform the original data into columns for each 500xy ID and assign the value from the one column.


Thank you, I will trying applying your suggestion. What is the difference between One-to-Many node and the One-to-Many PMML node?

The PMML Variant can store your rules and apply them on a new file

1 Like

The following entry explains the details of PMML in KNIME https://www.knime.com/blog/pmml-integration-in-knime


okay, thank you for all of your help

thank you very much for the information

This topic was automatically closed 182 days after the last reply. New replies are no longer allowed.