How to subtract a complete character vector with r

问题:

I want to subtract y from x, which means remove one "A", three "B" and one "E" from x, so xNew will be c("A", "C", "A","B","D"). It also means

length(xNew)=length(x) - length(y)
x <- c("A","A","C","A","B","B","B","B","D","E")
y <- c("A","B","B","B","E")

setdiff doesn't work because

xNew <- setdiff(x,y)
xNew 
[1] "C" "D"

match also doesn't work

xNew <- x[-match(y,x)]
xNew
[1] "A" "C" "A" "B" "B" "B" "D"

It removes "B" on the fifth position 3 times, so there are still three "B" left.

Is anyone know how to do this, is there a function available in R or we should write a private function? Thanks a lot in advance.

回答1:

You can use the function pmatch:

x[-pmatch(y,x)]
#[1] "A" "C" "A" "B" "D"

Edit
If your data can be strings of more than 1 character, here is an option to get what you want:

xNew <- unlist(sapply(x[!duplicated(x)], 
                      function(item, tab1, tab2) {
                          rep(item,
                              tab1[item] - ifelse(item %in% names(tab2), tab2[item], 0))
                       }, tab1=table(x), tab2=table(y)))

Example

x <- c("AB","BA","C","CA","B","B","B","B","D","E")
y <- c("A","B","B","B","E")
xNew
#  AB   BA    C   CA    B    D 
#"AB" "BA"  "C" "CA"  "B"  "D"