In “One to Many” node, include the “value” column and check the “Remove included columns from output” option. Then use a “GroupBy” node and select the ID column as the grouping column and for aggregation choose “Maximum” for all value columns (you can use “Type Based Aggregation”).
Thanks Armin, it works. One issue I have is that the data set has a high dimensionality (e.g. 100k rows, 5k columns). Therefore, the “one to many” node is time consuming. While I will use PCA/SVD afterwards to reduce the number of dimensions, I was wondering if there is a way to make it run faster.