testthat (version 2.2.1)

expect_known_output: Expectations: is the output or the value equal to a known good value?

Description

For complex printed output and objects, it is often challenging to describe exactly what you expect to see. expect_known_value() and expect_known_output() provide a slightly weaker guarantee, simply asserting that the values have not changed since the last time that you ran them.

Usage

expect_known_output(object, file, update = TRUE, ..., info = NULL,
  label = NULL, print = FALSE, width = 80)
expect_known_value(object, file, update = TRUE, ..., info = NULL,
  label = NULL, version = 2)
expect_known_hash(object, hash = NULL)

Arguments

object

Computation and value to compare it to.

Both arguments supports limited unquoting to make it easier to generate readable failures within a function or for loop. See quasi_label for more details.

file

File path where known value/output will be stored.

update

Should the file be updated? Defaults to TRUE, with the expectation that you'll notice changes because of the first failure, and then see the modified files in git.

...

For expect_equal() and expect_equivalent(), passed on compare(), for expect_identical() passed on to identical(). Used to control the details of the comparison.

info

Extra information to be included in the message. This argument is soft-deprecated and should not be used in new code. Instead see alternatives in quasi_label.

label

Used to customise failure messages. For expert use only.

print

If TRUE and the result of evaluating code is visible this will print the result, ensuring that the output of printing the object is included in the overall output

width

Number of characters per line of output. This does not inherit from getOption("width") so that tests always use the same output width, minimising spurious differences.

version

The serialization format version to use. The default, 2, was the default format from R 1.4.0 to 3.5.3. Version 3 became the default from R 3.6.0 and can only be read by R versions 3.5.0 and higher.

hash

Known hash value. Leave empty and you'll be informed what to use in the test output.

Details

These expectations should be used in conjunction with git, as otherwise there is no way to revert to previous values. Git is particularly useful in conjunction with expect_known_output() as the diffs will show you exactly what has changed.

Note that known values updates will only be updated when running tests interactively. R CMD check clones the package source so any changes to the reference files will occur in a temporary directory, and will not be synchronised back to the source package.

Examples

# NOT RUN {
tmp <- tempfile()

# The first run always succeeds
expect_known_output(mtcars[1:10, ], tmp, print = TRUE)

# Subsequent runs will succeed only if the file is unchanged
# This will succeed:
expect_known_output(mtcars[1:10, ], tmp, print = TRUE)

# }
# NOT RUN {
# This will fail
expect_known_output(mtcars[1:9, ], tmp, print = TRUE)
# }