My code is as below:

data = data.frame(x1 = c(1,1,1,1)
                 ,x2 = c(0,1,0,1)
                 ,x3 = c(1,1,0,1),x4 = c(1,1,0,0)) %>% rowSums

data%>%
        case_when(. == 0 ~ 0,
                  . %in% c(1,2)~ 1,
                  . %in% c(3:5)~ 2)

The sample data is as below:

where x1,x2,x3,x4 are in one data frame and they are binary variables. Then, the rowsums of x1,x2,x3,x4 are calculated.

The result is as below:

I would like to use case_when to do classification, however, when I run the above code, the error: ! Case 1 (.) must be a two-sided formula, not a double vector. also appears and I cannot solve it by using different method...

CodePudding user response：

The pipe inserts the left-hand expression as the first argument into the right-hand side call.

That is, your call is equivalent to:

case_when(data,
          data == 0 ~ 0,
          data %in% c(1,2) ~ 1,
          data %in% c(3:5) ~ 2)

To prevent this, surround the right-hand side with {…}:

data %>% {
  case_when(. == 0 ~ 0,
            . %in% c(1,2) ~ 1,
            . %in% c(3:5) ~ 2)
}

The documentation gives the following description:

For example, iris %>% subset(1:nrow(.) %% 2 == 0) is equivalent to iris %>% subset(., 1:nrow(.) %% 2 == 0) but slightly more compact. It is possible to overrule this behavior by enclosing the rhs in braces. For example, 1:10 %>% {c(min(.), max(.))} is equivalent to c(min(1:10), max(1:10)).