The first argument to every expect_
function can use unquoting to
construct better labels. This makes it easy to create informative labels when
expectations are used inside a function or a for loop. quasi_label()
wraps
up the details, returning the expression and label.
Arguments
- quo
A quosure created by
rlang::enquo()
.- label
An optional label to override the default. This is only provided for internal usage. Modern expectations should not include a
label
parameter.- arg
Argument name shown in error message if
quo
is missing.
Value
A list containing two elements:
- val
The evaluate value of
quo
- lab
The quasiquoted label generated from
quo
Limitations
Because all expect_
function use unquoting to generate more informative
labels, you can not use unquoting for other purposes. Instead, you'll need
to perform all other unquoting outside of the expectation and only test
the results.
Examples
f <- function(i) if (i > 3) i * 9 else i * 10
i <- 10
# This sort of expression commonly occurs inside a for loop or function
# And the failure isn't helpful because you can't see the value of i
# that caused the problem:
show_failure(expect_equal(f(i), i * 10))
#> Failed expectation:
#> f(i) (`actual`) not equal to i * 10 (`expected`).
#>
#> `actual`: 90.0
#> `expected`: 100.0
# To overcome this issue, testthat allows you to unquote expressions using
# !!. This causes the failure message to show the value rather than the
# variable name
show_failure(expect_equal(f(!!i), !!(i * 10)))
#> Failed expectation:
#> f(10) (`actual`) not equal to 100 (`expected`).
#>
#> `actual`: 90.0
#> `expected`: 100.0