how do you apply a fucntion to only numeric fields in data frame in R (excluding headers)

Question

I have a data frame this data frame:

    structure(list(Time = structure(1:3, .Label = c("1/13/15 12:14 PM", 
"1/13/15 13:14 PM", "1/13/15 14:14 PM"), class = "factor"), Server1 = structure(1:3, .Label = c("3", 
"5", "7"), class = "factor"), Server2 = structure(1:3, .Label = c("0", 
"1.3", "34"), class = "factor")), .Names = c("Time", "Server1", 
"Server2"), row.names = c(NA, -3L), class = "data.frame")

I need to globaly apply a rounding function to all data points, excluding headers and datetime formated data points:

I tried this:

sapply(df, function(x) round(df[,2:ncol(x)],2))

not working, any ideas?

Answer 1

Since you've got all factor columns there, you'll likely need to manually determine which ones are to be converted. Here's a function we can use to convert factors to their original numeric values. This uses the recommended method as described in the Warning section of help(factor).

f <- function(x) as.numeric(levels(x))[x]

Next we can apply this function to the desired columns and round them at the same time. In this case we want to apply f() to every column except the first, so we use the index [-1].

df[-1] <- lapply(df[-1], function(x) round(f(x), 2))
df
#               Time Server1 Server2
# 1 1/13/15 12:14 PM       3     0.0
# 2 1/13/15 13:14 PM       5     1.3
# 3 1/13/15 14:14 PM       7    34.0

Check the resulting column classes:

sapply(df, class)
#     Time   Server1   Server2 
# "factor" "numeric" "numeric"