Improve documentation around Regex-based test steps

[AsgerIversen]

Originally filed January 14 2022 by Rolf Madsen on GitLab

the SCPI and Process steps both uses regular expressions to do some simple parsing of the output, but actually the behavior is a bit surprising or at least not obvious.

Lets say the output of a SCPI step was:

"1.0, 2.0, 3.0"

In order to parse this, id expect this to be enough:

Regex: ([^,]+),?
Column Names: X, Y, Z
Behavior: Groups as Columns

This however gives an error saying the the number of columns does not match the number of groups. The regex for this means that I will instead get three rows with one column in each.

To do what I want I have to do like this:

Regex: ([^,]+),([^,]+),?([^,]+)
Column Names: X, Y, Z
Behavior: Groups as Columns

I think we should document the behavior with some examples or find more intuitive behaviors.