CV values cannot be pre-calculated when passing data into TwoRegression. This spares users from having to make the calculations themselves, but it also prevents them from doing so (e.g., if they want to test out a different approach to the sliding window calculations).
The current implementation either results in the existing values being overwritten after re-calculating the CV values, or else in adding a new column to the data with an essentially duplicated name. These aren't especially desirable behaviors, and resolving them could help with addressing #1.
It also shouldn't be too hard to test exists(cv_varname, AG). This could potentially be added as an optional check that, when enabled, would warn users that CV is not being calculated due to the presence of the existing variable.
CV values cannot be pre-calculated when passing data into
TwoRegression. This spares users from having to make the calculations themselves, but it also prevents them from doing so (e.g., if they want to test out a different approach to the sliding window calculations).The current implementation either results in the existing values being overwritten after re-calculating the CV values, or else in adding a new column to the data with an essentially duplicated name. These aren't especially desirable behaviors, and resolving them could help with addressing #1.
It also shouldn't be too hard to test
exists(cv_varname, AG). This could potentially be added as an optional check that, when enabled, would warn users that CV is not being calculated due to the presence of the existing variable.