Help with Time Series Data Windowing Based on Target Columns in KNIME

KNIME Analytics Platform
1

Hi everyone,

I am working with a time series dataset in KNIME that includes stock prices and various indicators (indicator1, indicator2, etc.). Each row in my dataset represents a different indicator and one time step (see screen shot below).

image
image1224×549 35.8 KB

I want to perform windowing on this data, where multiple time steps are consolidated into a single row based on selected target columns. Specifically, I need to:

  1. Control the window size.
  2. Perform the windowing in reverse order.

Below is an example of how I would like the data table to look after windowing:

image
image1370×534 68.6 KB

I am not sure how to approach this in KNIME. Can anyone guide me on how to achieve this?

Thank you in advance for your help!

2

Not sure if its the best solution, but something like this should work.

You could loop through the column renaming or copy and paste for each set of columns you need to rename.

You can play with it and sort the columns to get the exact format you’re looking for.

Input Data:
image

Output:
image

Regards,
Wali Khan
windows.knwf (23.2 KB)

1 Like