Passing function input to name dimnames-CodePudding

I am trying to assign names to list of dimnnames using inputs from an function, but the names only output the function input name (i.e. if function(test1 = "abc)" dimnames outputs "test1", not "abc"). An example function is here:

# example data
set.seed(123)
mcmr <- data.frame(type = rep(c("fm", "xp","mg"), each = 8),
                      teq = rep(c("string","ribbon"),  12),
                      value = sample(0:1, 24, replace = TRUE))

#example function
mcm_fun <- function(technique, test1, test2){
  test1 <- mcmr[mcmr$type == test1 & mcmr$teq == technique, "value"]
  test2 <- mcmr[mcmr$type == test2 & mcmr$teq == technique, "value"]
  matrix(c(table(test1), table(test2)), nrow = 2,
                dimnames = list(test1 = c("Positive", "Negative"), # want to name with test1 input
                                test2 = c("Positive", "Negative"))) # want to name with test2 input
}

Which outputs

mcm_fun("string", "mg", "xp")

#          test2
#test1      Positive Negative
#  Positive        1        3
#  Negative        3        1

The desired output would include the function input for test1 and test2:

#          mg
#xp      Positive Negative
#  Positive        1        3
#  Negative        3        1

I have tried assign() and get() but cant seem to get this to work and all I can seem to find is info on assigning the dimnames themselves.

CodePudding user response：

The labels you're asking about are the names of the dimnames of the table, so you can set them with:

names(dimnames(m)) <- nms

In your case, just add that assignment step into your function:

mcm_fun <- function(technique, t1, t2){
    test1 <- mcmr[mcmr$type == t1 & mcmr$teq == technique, "value"]
    test2 <- mcmr[mcmr$type == t2 & mcmr$teq == technique, "value"]
    m <- matrix(c(table(test1), table(test2)), nrow = 2,
           dimnames = list(test1 = c("Positive", "Negative"), # want to name with test1 input
                           test2 = c("Positive", "Negative"))) # want to name with test2 input
    names(dimnames(m)) <- c(t1, t2)
    return(m)
}

mcm_fun("string", "mg", "xp")
          xp
mg         Positive Negative
  Positive        3        1
  Negative        1        3

Note that I had to rename your arguments since you were overwriting them with the results of the test.

CodePudding user response：

The list cannot evaluate the objects on the lhs of =, we can use := with lst (from dplyr or purrr)

library(purrr)
mcm_fun <- function(technique, test1, test2){
  # assigned to a new object as the 
  # test1, test2 are assigned later to subset data
  nm1 <- test1
  nm2 <- test2
  test1 <- mcmr[mcmr$type == test1 & mcmr$teq == technique, "value"]
  test2 <- mcmr[mcmr$type == test2 & mcmr$teq == technique, "value"]
  matrix(c(table(test1), table(test2)), nrow = 2,
                dimnames = lst(!!nm1 := c("Positive", "Negative"), # want to name with test1 input
                                !!nm2 := c("Positive", "Negative"))) # want to name with test2 input

                            
}

-testing

> mcm_fun("string", "mg", "xp")
          xp
mg         Positive Negative
  Positive        3        1
  Negative        1        3