I am pretty new to R, about 3 months in, and when I was trying to run a regression R shot me this error, Error: unexpected input in "reg1 <- lm(_" the variable I use has an underscore, and some other variables too, I didn't know if R support underscore in a regression or not as thats the first time I had a variable with an underscore in it's name. If it doesn't, how can I change to name?

CodePudding user response：

As good practice, always begin variable/column names with letters (although not explicitly the rule, and you can technically start with the period symbol, this will save hassle). When dealing with data imported into R with predefined column names (or just when dealing with dataframes in general) you can rename columns in the dataframe df as follows

names(df)[names(df) == 'OldName'] <- 'NewName'

CodePudding user response：

If you really need to, you can protect 'illegal' names with back-quotes (although I agree with other answers/comments that this is not good practice ...)

dd <- data.frame(`_y`=rnorm(10), x = 1:10, check.names=FALSE)
names(dd)
## [1] "_y" "x" 
lm(`_y` ~ x, data = dd)