is rule engine node the right node?

I use RULE ENGINE NODE in order to detect and classify some rows.

Some of them are not classified, because the col dedicated for this (16_w_…) is empty.
In my mind I would like to do something like this:

If preditction have as result an empty cell, but others rows in a table have some col in common (ex. 13_…87JB AND 46_OM… 20 or 75), take the prediction you already discover for that rows.

In the attached file, you can see the first 8 raw are without prediction (because col 16 is empty), but U see, they have other in common (col 13 ad 46); I’d like to take the prediction of raw detected e copy the result if there are others in common.

I like to use the rule engine dictionary for things like this.

You could use the duplicate filter to narrow down to a list of unique values in the columns 13 / 16 / prediction, and then convert them into a formula that only applies them to blank cells.

You could also do this taking a join approach, by splitting your blank prediction rows into another stream and joining them back to get the referenced values required.

I uploaded both approaches in this thread if you want to see sample workflows. If you still need help, then upload a workflow with sample data.

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.