R code for if-then-do code blocks

Dear Paul,

R's power is that is works vectorised. Unlike SAS which is rowbased. Using
R in a SAS way will lead to very slow code.

Your examples can be written vectorised

d1 %>%
  rownames_to_column("ID") %>%
  mutate(
    test1 = ifelse(gender == "f" & workshop == 1, 7, 0),
    test2 = ifelse(gender == "f" & workshop == 1, test1 + 2, 0),
    test4 = ifelse(gender == "f" & workshop == 1, 1, 0),
    test5 = test4
  )

Here is a speed comparison.

library(microbenchmark)
microbenchmark(
  vector = {d1 %>%
    rownames_to_column("ID") %>%
    mutate(
      test1 = ifelse(gender == "f" & workshop == 1, 7, 0),
      test2 = ifelse(gender == "f" & workshop == 1, test1 + 2, 0),
      test4 = ifelse(gender == "f" & workshop == 1, 1, 0),
      test5 = test4
    ) },
  rowbased = {d1 %>%
  rownames_to_column("ID") %>%
  mutate(test1 = NA, test2 = NA, test4 = NA, test5 = NA) %>%
  ddply("ID",
        within,
        if (gender == "f" & workshop == 1) {
          test1 <- 1
          test1 <- 6 + test1
          test2 <- 2 + test1
          test4 <- 1
          test5 <- 1
        } else {
          test1 <- test2 <- test4 <- test5 <- 0
        })}
)


Best regards,

Thierry

ir. Thierry Onkelinx
Statisticus / Statistician

Vlaamse Overheid / Government of Flanders
INSTITUUT VOOR NATUUR- EN BOSONDERZOEK / RESEARCH INSTITUTE FOR NATURE AND
FOREST
Team Biometrie & Kwaliteitszorg / Team Biometrics & Quality Assurance
thierry.onkelinx at inbo.be
Havenlaan 88 bus 73, 1000 Brussel
www.inbo.be

///////////////////////////////////////////////////////////////////////////////////////////
To call in the statistician after the experiment is done may be no more
than asking him to perform a post-mortem examination: he may be able to say
what the experiment died of. ~ Sir Ronald Aylmer Fisher
The plural of anecdote is not data. ~ Roger Brinner
The combination of some data and an aching desire for an answer does not
ensure that a reasonable answer can be extracted from a given body of data.
~ John Tukey
///////////////////////////////////////////////////////////////////////////////////////////

<https://www.inbo.be>


Op ma 17 dec. 2018 om 16:30 schreef Paul Miller via R-help <
r-help at r-project.org>:

Hello All,

Season's greetings!

 Am trying to replicate some SAS code in R. The SAS code uses if-then-do
code blocks. I've been trying to do likewise in R as that seems to be the
most reliable way to get the same result.

Below is some toy data and some code that does work. There are some things
I don't necessarily like about the code though. So I was hoping some people
could help make it better. One thing I don't like is that the within
function reverses the order of the computed columns such that test1:test5
becomes test5:test1. I've used a mutate to overcome that but would prefer
not to have to do so.

 Another, perhaps very small thing, is the need to calculate an ID
variable that becomes the basis for a grouping.

I did considerable Internet searching for R code that conditionally
computes blocks of code. I didn't find much though and so am wondering if
my search terms were not sufficient or if there is some other reason. It
occurred to me that maybe if-then-do code blocks like we often see in SAS
as are frowned upon and therefore not much implemented.

I'd be interested in seeing more R-compatible approaches if this is the
case. I've learned that it's a mistake to try and make R be like SAS. It's
better to let R be R. Trouble is I'm not always sure how to do that.

Thanks,

Paul


d1 <- data.frame(workshop=rep(1:2,4),
                gender=rep(c("f","m"),each=4))

library(tibble)
library(plyr)

d2 <- d1 %>%
  rownames_to_column("ID") %>%
  mutate(test1 = NA, test2 = NA, test4 = NA, test5 = NA) %>%
  ddply("ID",
        within,
        if (gender == "f" & workshop == 1) {
          test1 <- 1
          test1 <- 6 + test1
          test2 <- 2 + test1
          test4 <- 1
          test5 <- 1
        } else {
          test1 <- test2 <- test4 <- test5 <- 0
        })

        [[alternative HTML version deleted]]