R：在将所有元素粘贴到单个字符串之前，最优雅的方法来清理数据框

Question

I have a data frame similar to below:

df.example <- data.frame(c(2:3, NA, "John"), c("Mary", NA, "", NA), c("Dennis", 4:6))

 2     Mary     Dennis
 3     John     4
 <NA>           5      # note the empty string
 John  <NA>     6      # "John" occurs twice

I have omitted the row and column names because they contain no useful information.

I would like to convert this to a 1D list, containing ALL values in the data frame, removing NA and empty string values. I would like to retain only unique values, and concatenate them so I can use them as an argument for a SQL query later in the code.

This is my current solution:

temp <- df.example[!is.na(df.example)]

 # "2" "3" "John" "Mary" "John" "" "Dennis" "4" "5" "6"     

temp <- unique(temp[temp != ""])

 # "2" "3" "John" "Mary" "Dennis" "4" "5" "6"     

output <- paste0("'",temp,"'", collapse = ",")

 #"'2','3','John','Mary','Dennis','4','5','6'"

Now I can pass this string to SQL.

This seems like way too many lines of code to do what I was expecting to be a one-liner. Is there a more elegant way?

Answer 1

I agree with Rich Scriven's comment. However, if you want a one-liner answer perhaps something like:

df.example <- data.frame(c(2:3, NA, "John"), c("Mary", NA, "", NA), c("Dennis", 4:6))
names(table(as.vector(t(df.example)),exclude=c("",NA)))
    # [1] "2"      "3"      "4"      "5"      "6"      "Dennis" "John"   "Mary"