(object 'nLev' not found) on VIM regressionImp function

[varvanih]

Hello, I am trying to use VIM regressionImp function. But I encounter the below error when trying to run the code:

Error in regressionImp_work(formula = formula, data = data, family = family, :
object 'nLev' not found

This is what I am trying to run:

---

rm(list = ls())

library("VIM")

library("readxl")

setwd("E:\data_sets")

dataincomplete <- read_excel("4-gauss.xlsx")

form1 <- X1 ~ X3

for (v in c("X1")) {
dataincomplete[sample(1:nrow(dataincomplete), round(nrow(dataincomplete) / 10)), v] <- NA
}

dataimputed <- regressionImp(form1, dataincomplete, family = "AUTO", robust = TRUE, mod_cat = FALSE)

I attached my data file to this message.

My R version is 3.6.1 run on Windows 10 64bit.

Any help is really appreciated.

Thanks,
Farahani

4-gauss.xlsx

[matthias-da]

I corrected the code. Thanks for pointing this error.

Please reinstall VIM from GitHub with the devtools package.

library(devtools)
install_github("statistikat/VIM")

Best
Matthias

[varvanih]

Hello Matthias

Thank you very much for your help. That worked for me.

Also I realized that when I read the data set from an Excel file, its class is not simply "data.frame" but {"tbl_df" "tbl" "data.frame"}.

So I used below line to change its class to "data.frame" and this also solved the problem and now it works also with your previous VIM version:

DATA = as.data.frame(DATA)

Regards,
Farahani

[chrhennig]

Still produces an error for me. Apparently regressionImp gets into trouble with variables of integer type. Converting them to numeric makes the problem go away for me.