R.utils (version 2.12.3)

captureOutput: Evaluate an R expression and captures the output

Description

Evaluate an R expression and captures the output.

Usage

captureOutput(expr, file=NULL, append=FALSE, collapse=NULL, envir=parent.frame())

Value

Returns captured output as a character

vector.

Arguments

expr

The R expression to be evaluated.

file

A file name or a connection to where the output is directed. Alternatively, if NULL the output is captured to and returned as a character vector.

append

If TRUE, the output is appended to the file or the (unopened) connection, otherwise it overwrites.

collapse

A character string used for collapsing the captured rows. If NULL, the rows are not collapsed.

envir

The environment in which the expression is evaluated.

Author

Henrik Bengtsson

Details

This method imitates capture.output with the major difference that it captures strings via a raw connection rather than via internal strings. The latter becomes exponentially slow for large outputs [1,2].

References

[1] R-devel thread 'capture.output(): Using a rawConnection() [linear] instead of textConnection() [exponential]?', 2014-02-04. https://stat.ethz.ch/pipermail/r-devel/2014-February/068349.html [2] JottR blog post 'PERFORMANCE: captureOutput() is much faster than capture.output()', 2015-05-26. https://www.jottr.org/2014/05/26/captureoutput/

See Also

Internally, eval() is used to evaluate the expression. and capture.output to capture the output.

Examples

Run this code
# captureOutput() is much faster than capture.output()
# for large outputs when capturing to a string.
for (n in c(10e3, 20e3, 30e3, 40e3)) {
  printf("n=%d\n", n)

  x <- rnorm(n)

  t0 <- system.time({
    bfr0 <- capture.output(print(x))
  })
  print(t0)

  t1 <- system.time({
    bfr <- captureOutput(print(x))
  })
  print(t1)
  print(t1/t0)

  bfr2n <- captureOutput(print(x), collapse="\n")
  bfr2r <- captureOutput(print(x), collapse="\r")

  stopifnot(identical(bfr, bfr0))
} # for (n ...)

Run the code above in your browser using DataLab