get

`get()` retrieves objects by name. ```{r} x <- 1:5 get("x") ``` You can use it with functions too, but often you want [`match.fun()`](https://www.rdocumentation.org/packages/base/topics/match.fun) for that task. ```{r} get("var") match.fun("var") ``` You can specify where to look for the object. ```{r} get("abs", as.environment("package:base")) ``` If we look in an environment higher up the [`search()`](https://www.rdocumentation.org/packages/base/topics/search) path then get will still find what we asked for. ```{r} get("abs", as.environment("package:stats")) ``` …unless we specify `inherits = FALSE` ```{r} tryCatch( get("abs", as.environment("package:stats"), inherits = FALSE), error = print ) ``` It is a good idea to check to see if the variable [`exists()`](https://www.rdocumentation.org/packages/base/topics/exists) before getting it. ```{r} if(exists("abs", as.environment("package:stats"), inherits = FALSE)) { get("abs", as.environment("package:stats"), inherits = FALSE) } else { message("Could not find abs in the stats package.") } ``` `get0()` provides a more efficient alternative to the previous code pattern. ```{r} abs_fn <- get0("abs", as.environment("package:stats"), inherits = FALSE) if(is.null(abs_fn)) { message("Could not find abs in the stats package.") } ``` `mget()` is useful for retrieving objects satisfying some criterion from an `environment`. For example, you can get all primitive functions from the `methods` package. ```{r} methods_env <- as.environment("package:methods") methods_var_names <- ls(methods_env) methods_vars <- mget(methods_var_names, methods_env) (primitive_fns_in_methods <- Filter(is.primitive, methods_vars)) ``` Another use of `mget()` is for grouping related variables into a `list`. Suppose you have a few variables that you want to calculate the mean of. Working with individual variables can be tiresome. ```{r} x1 <- 1:5 x2 <- (1:5) ^ 2 x3 <- (1:5) ^ 3 x4 <- (1:5) ^ 4 x5 <- (1:5) ^ 5 # The following code involves cutting and pasting, which makes it error-prone. mean(x1) mean(x2) mean(x3) mean(x4) mean(x5) # By using a list, you can use sapply to loop over elements, and save some repeated typing. x_var_names <- ls(globalenv(), pattern = "^x[1-5]$") x_list <- mget(x_var_names, globalenv()) sapply(x_list, mean) ``` `mget()` has an `ifnotfound` argument that saves you having to do if-exists-then-get. ```{r} mget( c("var", "abs"), as.environment("package:stats"), ifnotfound = NA, inherits = FALSE ) ``` `dynGet()` works its way back up the call stack until it finds the variable that you are looking for. ```{r} f <- function(x) { z <- 3 g(x + 1, z - 1) } g <- function(x, y) { message('dynGet("x") = ', dynGet("x")) message('dynGet("y") = ', dynGet("y")) message('dynGet("z") = ', dynGet("z")) message( 'dynGet("<?>") = ', dynGet("<?>", ifnotfound = "did not find <?>") ) } f(0) ```