Try an Expression Allowing Error Recovery

Description

try is a wrapper to run an expression that might fail and allow the user's code to handle error-recovery.

Usage

try(expr, silent = FALSE,
    outFile = getOption("try.outFile", default = stderr()))

Arguments

Details

try evaluates an expression and traps any errors that occur during the evaluation. If an error occurs then the error message is printed to the stderr connection unless options("show.error.messages") is false or the call includes silent = TRUE. The error message is also stored in a buffer where it can be retrieved by geterrmessage. (This should not be needed as the value returned in case of an error contains the error message.)

try is implemented using tryCatch; for programming, instead of try(expr, silent = TRUE), something like tryCatch(expr, error = function(e) e) (or other simple error handler functions) may be more efficient and flexible.

It may be useful to set the default for outFile to stdout(), i.e.,

  options(try.outFile = stdout()) 

instead of the default stderr(), notably when try() is used inside a Sweave code chunk and the error message should appear in the resulting document.

Value

The value of the expression if expr is evaluated without error, but an invisible object of class "try-error" containing the error message, and the error condition as the "condition" attribute, if it fails.

See Also

options for setting error handlers and suppressing the printing of error messages; geterrmessage for retrieving the last error message. The underlying tryCatch provides more flexible means of catching and handling errors.

assertCondition in package tools is related and useful for testing.

Examples

## this example will not work correctly in example(try), but
## it does work correctly if pasted in
options(show.error.messages = FALSE)
try(log("a"))
print(.Last.value)
options(show.error.messages = TRUE)

## alternatively,
print(try(log("a"), TRUE))

## run a simulation, keep only the results that worked.
set.seed(123)
x <- stats::rnorm(50)
doit <- function(x)
{
    x <- sample(x, replace = TRUE)
    if(length(unique(x)) > 30) mean(x)
    else stop("too few unique points")
}
## alternative 1
res <- lapply(1:100, function(i) try(doit(x), TRUE))
## alternative 2
## Not run: res <- vector("list", 100)
for(i in 1:100) res[[i]] <- try(doit(x), TRUE)
## End(Not run)
unlist(res[sapply(res, function(x) !inherits(x, "try-error"))])