r - Why na.omit adds a row to empty dataframe? -

i have code doing like

d <- load data.frame, possible empty one... d <- na.omit(d) if (nrow(d)>0) {  something... } 

this appears wrong due fact applying na.omit empty data frame adds row it:

data.frame()   data frame 0 columns , 0 rows  na.omit(data.frame())   data frame 0 columns , 1 rows 

why na.omit doing me?

this because na.omit has logical vector omit (in code) gets set false rows keep , true rows remove.

however, omit set false before checking on input data.frame rows remove , update value of omit. since input empty data.frame, there no values updated , omit retains earlier set false. , then, na.omit calls:

object[!omit, , drop=false] 

which in case is:

data.frame()[true, , drop=false] 

which gives:

# data frame 0 columns , 1 rows 

---

here's code na.omit.data.frame (which can obtained doing gets3method("na.omit", "data.frame")). part not run empty data.frame commented out.

n <- length(object) omit <- false vars <- seq_len(n) # equals integer(0) in case (j in vars) { # loop not run @ #     x <- object[[j]] #     if (!is.atomic(x))  #         next #     x <- is.na(x) #     d <- dim(x) #     if (is.null(d) || length(d) != 2l)  #         omit <- omit | x #     else (ii in 1l:d[2l]) omit <- omit | x[, ii] # } xx <- object[!omit, , drop = false] # if (any(omit > 0l)) { # not run #     temp <- setnames(seq(omit)[omit], attr(object, "row.names")[omit]) #     attr(temp, "class") <- "omit" #     attr(xx, "na.action") <- temp # } xx 

---

solution:

you use complete.cases instead.

df <- data.frame() df[complete.cases(df), ] # data frame 0 columns , 0 rows  df <- data.frame(x=1:2, y=c(2,na)) df[complete.cases(df), ] #   x y # 1 1 2