Add column with fixed value

[Dr_Van_Nostrand]

Hi,

Is there an easy way to add a column with the same value for all rows? Is it possible to use a flow variable for setting the value?

Thanks in advance!

[tobias.koetter]

Hello,

you can use the "Variable To TableColumn" node to add the value of a flow variable as column to a data table.

Bye,

Tobias

[richards99]

You can use the flow variable (as described above), or you can use the Rule Engine node, or you can use the Maths node too by just entering a number in to the Maths Expression box.

If you already have the flow variable, you can define this in the Maths Expression too in the Maths node.

 

Simon.

[Ellert_van_Koperen]

I usualy use the simple java node for this.
It will just contain something like

return(6);

[Dr_Van_Nostrand]

Thanks for all tips. I fixed it with the Rule Engine node.

Cheers!

[Iris]

To give everyone an update to this one.

We have a dedicated node, the constant value node which can do this very nicely.

Constant Value Column

[amarocchino]

That seems terrific, but I am stucked when I’d like to extract coefficients from a linear regression (node: linear regression learner) and add to the original dataframe a column with the intercept value and one column with the gradient value. Is there a node to do so?

[ScottF]

Not 100% sure what you’re trying to do, but if you attach a Pivoting node to the bottom port of the Linear Regression Learner, and configure it with no grouping, pivot = Variable, and manual aggregation = mean(Coeff) you will get a row corresponding to the coefficients from the regression.

You could then use the Cross Joiner to append the coefficients row-wise back to the original data. But then again I may be misunderstanding you, so feel free to clarify if that’s the case.

[amarocchino]

Hi Scott, you had interpreted it correctly. I needed to extract Linear Regression Coefficients and append them to the original table.

To give you a more general context, this is what I am trying to do. I have data than need to be either positive or negative. I use a k-means (cluster number=2) to check whether the two centroids have opposite sign. If yes, it passes a first test, at this point I use a linear regression for the positive values, first, and for the negative values, secondly. I am checking if the linear regression coefficients are in a given range.

If you have any suggestions or comments, please feel free to let me know I’ll be happy to improve my pipeline.