检索 rData 文件中所有 data.frame class 对象的路径

Question

I have multiple.rData files whose top level Global Environment variables are a mix of data.frames, lists, deeply nested lists. I know that many of the nested lists have within them data.frame types, but I'm having trouble retrieving the path to them.

I had a faced a similar problem before with another type of class using the following code

names(rapply(mget(ls(.GlobalEnv), envir=.GlobalEnv), length, classes="fluor.spectral.data", how="unlist"))

and while not the most elegant solution, it achieved what I needed and quickly. returning names like "Fluor.Spec.WA.M12.SC.13" which then allows me to manipulate the object after formating the '.' into '$'.

Can someone help me retrieve the path to all data.frame class types, nested or otherwise in highly variable.rData files? Thanks in advance

Answer 1

If you want to return all data.frames loaded in the global environment, either present as individual object or as element of a nested list, use rrapply in the rrapply -package (extension of base rrapply ).

library(rrapply)

w <- data.frame(1)
x <- list(1, 2, 3)
y <- 5
z <- list(1, 2, list(1, df = data.frame(a = 1, b = 2)))

rrapply(as.list(.GlobalEnv), classes = "data.frame", how = "flatten")
#> $w
#>   X1
#> 1  1
#> 
#> $df
#>   a b
#> 1 1 2

Setting classes = "data.frame" avoids recursion into data.frame columns (as base rapply would do), and how = "flatten" will return the collected data.frames as a flattened list.

NB: If you want to return the complete object paths to the found data.frames, set how = "prune" instead of how ="flatten" :

rrapply(as.list(.GlobalEnv), classes = "data.frame", how = "prune")
#> $w
#>   X1
#> 1  1
#> 
#> $z
#> $z[[1]]
#> $z[[1]]$df
#>   a b
#> 1 1 2

---

Edit: In order to also return data.frames present in slots of some S4-class, a possible way to extend the above call would be:

## define S4-class with a data.frame in "df" slot
userClass <- setClass("user", slots = c(df = "data.frame"))
v <- userClass(df = data.frame(user = 1))

rrapply(as.list(.GlobalEnv), 
                   classes = c("data.frame", "user"), 
                   f = function(x) {
                     if(class(x) == "user") {
                       slot(x, "df")
                     } else {
                       x
                     }
                   },
                   how = "flatten")
#> $v
#>   user
#> 1    1
#> 
#> $w
#>   X1
#> 1  1
#> 
#> $df
#>   a b
#> 1 1 2

In this case, classes = c("data.frame", "user") will check for data.frames and S4-objects of class "user" . The f function applied to the object, returns the object itself if it is a data.frame or the "df" slot if it is an S4-object.

Note that this code assumes that the S4-class name is known as well as the slot(s) which contain the data.frame objects.