Bug with variable names

[StephenGerry]

It seems like there is a bug which throws an error from the lms() function when the x variable name ends in a y (and the y variable does not begin with y).

As an example:

rm(list = ls())

library(gamlss)

data(abdom)

abdom$xy <- abdom$x
abdom$p <- abdom$y

m1 <- lms(p,xy, data=abdom, n.cyc=30)

This produces the error

Error in get(xvar, envir = as.environment(DaTa)) : object 'xp' not found

[zeileis]

I can confirm that this happens and don't understand where exactly it comes from. The data handling in the main lms() function works fine but the error occurs in the calibration() step. I didn't dig deeper from there because I hope that Mikis @mstasinopoulos can say more about what is happening.

[mstasinopoulos]

I look there is a problem.
It is due to the fact that we try to change the final fitted GAMLSS object to have the same form as the function of the `lms' call.
I was doing this before the call to plot for calibration and that was creating the problem

I move the call for calibration before the changes occur and it seems to work

The new lms is pushed in githab and it will work with new version

[StephenGerry]

Thanks both. That's good to hear that I wasn't going mad!

Great package by the way - I have used it a lot.

[zeileis]

Great! Glad the package is useful for you. If you want to try out whether Miki's update fixed your problem, you can install the development version of the package via:

install.packages("gamlss", repos = "https://gamlss-dev.R-universe.dev")