CODEWITHSUNDEEP
rreflectionmetaprogrammingcomputer-science
wespiserA
wespiserA

Reputation: 3168

does the R programming language have reflection?

Does R have reflection?

http://en.wikipedia.org/wiki/Reflection_(computer_programming)

basically what i want to do is this:

currentRun = "run287"
dataFrame$currentRun= someVar;

such that dataFrame$currentRun is equivalent to dataFrame$run287.
This hasn't prevented me from solving any problems, but from an academic standpoint I would like to know if R supports reflective programming. If so, how would one go about using reflection in the example given? Thank you!

Upvotes: 2

Views: 2555

Answers (3)

IRTFM
IRTFM

Reputation: 263451

The use of the "$" operator should be discouraged in programming because it does not evaluate its argument, unlike the more general "[["

currentRun = "run287"
dataFrame[[currentRun]]= someVar   # and the";" is superflous

> dat <- data.frame(foo = rnorm(10), bar = rnorm(10))
>  myVar <- "bar2"
> bar2 <- 1:10
>  dat[[myVar]] <- bar2
> str(dat)
'data.frame':   10 obs. of  3 variables:
 $ foo : num  -1.43 1.7091 1.4351 -0.7104 -0.0651 ...
 $ bar : num  -0.641 -0.681 -2.033 0.501 -1.532 ...
 $ bar2: int  1 2 3 4 5 6 7 8 9 10

Which will succeed if the properties (in particular length) of myVar are correct. It would not be correct to say the datFrame$currentRun is equivalent to dataFrame$run287, but is is correct that character variables can be interpreted as column names. There is also a eval(parse(text="...")) construct, but it is better to avoid if possible.

Upvotes: 6

kohske
kohske

Reputation: 66902

yes, R supports reflective programming.

here is an R version of the example:

foo <- function()1

# without reflection
foo()

# with reflection
get("foo")()

probably such as get, assign, eval is relevant. see online helps of them.

Upvotes: 9

Chase
Chase

Reputation: 69221

I'm not sure I fully grasped the wikipedia article, but does indexing with [ achieve your desired result? A trivial example:

> dat <- data.frame(foo = rnorm(10), bar = rnorm(10))
> myVar <- "bar"
> dat[ , myVar]
 [1]  1.354046574  0.551537607  0.779769817  0.546176894 -0.194116973  0.959749309
 [7] -1.560839187 -0.024423406 -2.487539955 -0.201201268

or

> dat[ , myVar, drop = FALSE]
            bar
1   1.354046574
2   0.551537607
3   0.779769817
....

Upvotes: 1

Related Questions