在数据框中添加字符串列（行）

Question

I have a data frame with columns that are all strings. There are occasional NA 's as well. I want to add all the columns into one vector row-by-row so that I can get only the unique strings in the entire data frame. Given a sample data like this:

        L1     L2    L3    L4
1      bed bicyle  <NA>  bird
2     <NA>   road   cat dance
3      cat   rock table   bed
4 mountain    bed clock  <NA>

Then I want something like this:

                                                     Total
1 bed bicyle bird road cat dance rock table mountain clock

I can obviously do this with a loop, but I am trying to figure out an apply or plyr type solution (which is more R -like). So far I can get a list of arrays but I then get stuck trying to add those arrays together. Is my approach even correct? Any help is appreciated.

df <- data.frame(L1 = c("bed", NA, "cat", "mountain"), 
                 L2 = c("bicyle", "road", "rock", "bed"), 
                 L3 = c(NA, "cat", "table", "clock"), 
                 L4 = c("bird", "dance", "bed", NA))
df <- lapply(df, unique)

Answer 1

You can simply do this:

unique(as.vector(t(df)))

Or, whitout NAs :

unique(na.omit(as.vector(t(df))))

Answer 2

Here's one way to accomplish your task:

require(dplyr)
df <- data.frame(L1 = c("bed", NA, "cat", "mountain"), 
                 L2 = c("bicyle", "road", "rock", "bed"), 
                 L3 = c(NA, "cat", "table", "clock"), 
                 L4 = c("bird", "dance", "bed", NA), 
                 stringsAsFactors = FALSE)

df <- df %>% unlist() %>% na.omit() %>% unique()

Output:

 [1] "bed"      "cat"      "mountain" "bicyle"   "road"     "rock"     "table"    "clock"    "bird"     "dance"