Error in `[.data.frame`(data, , var) : undefined columns selected (despite trying all the solutions suggested here)

[daniela-pazpez]

As other people, I am getting this message. I have tried using the 'multigroup2' function, but it created more issues. My code is:

pmodel3 <- psem (
lm(WORKE ~ cent_prestige + cent_inclusion + YEAR, merit_educ_na),
lm(PERMEAB ~ cent_prestige + cent_inclusion + YEAR, merit_educ_na),
lm(cent_prestige ~ cent_inclusion + YEAR, merit_educ_na))

multigroup(pmodel3, group = "education")

How can I fix this?

[ekyoungs]

I have the same problem.

Example code:

psem1.0 <- psem(
lm(ndvi200 ~ canopy200, data = data3),
lm (totHerb ~ ndvi200, data = data3),
lm(tot_nonHerb ~ totHerb, data = data3),
data = data3
)

summary(psem1.0) # works, everything looks normal to me

multigroup (psem1.0, group = 'zoning') # fails with the following error:

Error in [.data.frame(data, , var) : undefined columns selected

Looking at the psem object, 'zoning' is clearly stored in psem1.0$data
The error occurs whether 'zoning' is a factor or an integer.
All packages are up to date, and this occurs when I use the CRAN version or the development version of piecewiseSEM v2.3.0.
The meadow example runs fine, and I'm struggling to figure out why it can find 'grazing' but my psem can't find 'zoning'

If I try to direct it to 'zoning' more explicitly, i.e.,
multigroup (psem1.0, group = psem1.0$data$zoning)
I get:
Error in eval(predvars, data, env) : object 'greenway' not found # (greenway is a level of 'zoning')

Or, for the integer version of the grouping variable:
multigroup (psem1.0, group = psem1.0$data$zoning_integer)
Error in terms.formula(tmp, simplify = TRUE) :
invalid model formula in ExtractVars

I'd be happy to provide data (I'm not sure how to make useful fake data though)

Also tried multigroup2 (and its modification) from #284 but also with errors - this function seems not to work even on the meadows dataset in the documentation?

[LauraCappelatti]

I have the same issue when I include categorical variables (as factors). If I keep only continuous variables, then it works. And the model with categorical variables works without the multigroup. The multigroup2 that a user created caused me other problems.