remove rows which total 0 from large R data frame, where first and last columns are not numeric

Question

I have a large dataframe and need to remove a few rows from it as rowSums shows these to be 0 and thus impacting my downstream analysis. I've seen that you can do df[rowSums(df[, -1])>0, ] to achieve this, but I get an error stating that "x must be numeric" as both the first and last columns are not numeric. Any ideas of how to get around this?

Answer 1

We need to remove the first and last columns by indexing

i1 <- sapply(df, is.numeric) # // get a logical index for numeric columns
df[rowSums(df[i1]) > 0,]