tests/testthat/_snaps/cast.md

In r-lib/vctrs: Vector Helpers

Casting to named argument mentions 'match type '

Code
  vec_cast(1, "", x_arg = "foo", to_arg = "bar")
Condition
  Error:
  ! Can't convert `foo` <double> to match type of `bar` <character>.

Code
  vec_cast(1, "", x_arg = "foo")
Condition
  Error:
  ! Can't convert `foo` <double> to <character>.

cast errors create helpful messages (#57, #225)

Code
  vec_cast(1.5, 10L)
Condition
  Error:
  ! Can't convert from `1.5` <double> to <integer> due to loss of precision.
  * Locations: 1

Code
  vec_cast(factor("foo"), 10)
Condition
  Error:
  ! Can't convert `factor("foo")` <factor<c1562>> to <double>.

Code
  x <- tibble(a = tibble(b = 1.5))
  y <- tibble(a = tibble(b = 10L))
  vec_cast(x, y)
Condition
  Error:
  ! Can't convert from `x$a$b` <double> to `a$b` <integer> due to loss of precision.
  * Locations: 1

Code
  x <- tibble(a = tibble(b = factor("foo")))
  y <- tibble(a = tibble(b = 10))
  vec_cast(x, y)
Condition
  Error:
  ! Can't convert `x$a$b` <factor<c1562>> to match type of `a$b` <double>.

Code
  x <- tibble(a = tibble(b = factor("foo")))
  y <- tibble(a = tibble(b = 10))
  vec_cast_common(x, y)
Condition
  Error:
  ! Can't combine `..1$a$b` <factor<c1562>> and `..2$a$b` <double>.

vec_cast() only attempts to fall back if to is a data frame (#1568)

Code
  (expect_error(vec_cast(foobar(mtcars), 1), class = "vctrs_error_incompatible_type")
  )
Output
  <error/vctrs_error_cast>
  Error:
  ! Can't convert `foobar(mtcars)` <vctrs_foobar> to <double>.

can signal deprecation warnings for lossy casts

Code
  (expect_warning(expect_true(lossy_cast())))
Output
  <warning/lifecycle_warning_deprecated>
  Warning:
  Coercion with lossy casts was deprecated in vctrs 0.2.0.
  i Please use `allow_lossy_cast()` instead.
  i We detected a lossy transformation from `x` <fct> to `to` <fct>. The result will contain lower-resolution values or missing values. To suppress this warning, wrap your code with `allow_lossy_cast()`.

r-lib/vctrs documentation built on Oct. 30, 2024, 8:54 a.m.