Column in dataframe turns from integer to numeric when replacing NA with 0

In a dataframe, when I replace NA with 0 using obt[is.na(obt)] <- 0, the column type changes from integer to numeric.

str(obt)
 ...
 $ Units: int  NA NA 2 1 NA NA NA 1 NA NA ...

obt[is.na(obt)] <- 0

str(obt)
 ...
 $ Units: num  0 0 2 1 0 0 0 1 0 0 ...

How can I avoid that?

Sample data:

obt <- structure(list(Date = structure(c(19677, 19678, 19679, 19680, 
19681, 19682, 19683, 19684, 19685, 19686), class = "Date"), Title = structure(c(NA, 
NA, 3L, 3L, NA, NA, NA, 3L, NA, NA), levels = c("A", "D", "L", 
"C"), class = "factor"), Units = c(NA, NA, 2L, 1L, NA, NA, NA, 
1L, NA, NA)), row.names = c(NA, 10L), class = "data.frame")

>Solution :

Using L to indicate the number is an integer (same as using as.integer(0)).

obt$Units[is.na(obt$Units)] <- 0L

class(obt$Units)
[1] "integer"

Another example, when using dput

dput(as.integer(0))
0L

dput(as.numeric(0))
0

# R is using numeric when not explicitly stated (exceptions are e.g. sequences)
dput(0)
0