使用列表列或嵌套 data.frame 测试 tibbles 的相等性

Question

Tibbles（来自 tidyverse）可以包含列表列，这对于包含例如data.frame.

中传统上找不到的嵌套数据框或对象

这是一个例子：

library("dplyr")

nested_df <-
      iris %>%
      group_by(Species) %>%
      tidyr::nest() %>%
      mutate(model = purrr::map(data, lm, formula = Sepal.Length ~ .))

nested_df
#  # A tibble: 3 x 3
#   Species    data              model   
#   <fct>      <list>            <list>  
# 1 setosa     <tibble [50 × 4]> <S3: lm>
# 2 versicolor <tibble [50 × 4]> <S3: lm>
# 3 virginica  <tibble [50 × 4]> <S3: lm>

我正在用 testthat 编写一些测试：如何测试这些 data.frame 之间的相等性？

testthat::expect_equal 不起作用，因为 all.equal 和 dplyr::all_equal 都失败了：

all.equal(nested_df, nested_df)
# Error in equal_data_frame(target, current, ignore_col_order = ignore_col_order,  : 
#  Can't join on 'data' x 'data' because of incompatible types (list / list)

我考虑过使用 testthat::expect_true(identical(...))，但它通常过于严格。例如，定义完全相同的 nested_df2 不足以传递 identical，因为 lm 模型中嵌入的 terms 的 .Environment 属性不同，尽管模型相等并通过 all.equal.

identical(nested_df, nested_df2)
# [1] FALSE
identical(nested_df$model, nested_df2$model, ignore.environment = TRUE)
# [1] FALSE
all.equal(nested_df$model, nested_df2$model, tolerance = 0)
# [1] TRUE

如何测试 tibbles 与 nested_df 等列表列的相等性？

Answer 1

有点生硬的方法，但它似乎适用于您的示例：

all.equal.list(nested_df, nested_df)

# [1] TRUE

all.equal.list(nested_df, mutate(nested_df, Species = sample(Species)))

# [1] "Component “Species”: 2 string mismatches"

Answer 2

要扩展@utubun 的答案，您可以将 all.equal.list 包装在类似 testthat 的 expect_* 函数中：

expect_equal_tbl <- function(object, expected, ..., info = NULL) {
  act <- testthat::quasi_label(rlang::enquo(object), arg = "object")
  exp <- testthat::quasi_label(rlang::enquo(expected), arg = "expected")

  # all.equal.list is slightly problematic: it returns TRUE for match, and
  # returns a character vector when differences are observed. We extract
  # both a match-indicator and a failure message

  diffs <- all.equal.list(object, expected, ...)
  has_diff <- if (is.logical(diffs)) diffs else FALSE
  diff_msg <- paste(diffs, collapse = "\n")

  testthat::expect(
    has_diff,
    failure_message = sprintf(
      "%s not equal to %s.\n%s", act$lab, exp$lab, diff_msg
    ),
    info = info
  )

  invisible(act$val)
}

expect_equal_tbl(nested_df, nested_df, info = "YAY!")

expect_equal_tbl(nested_df, nested_df[1, ], info = "FAIL!")
 Error: `nested_df` not equal to nested_df[1, ].
Attributes: < Component “row.names”: Numeric: lengths (3, 1) differ >
Component “Species”: Lengths: 3, 1
Component “Species”: Lengths (3, 1) differ (string compare on first 1)
Component “data”: Length mismatch: comparison on first 1 components
Component “model”: Length mismatch: comparison on first 1 components
FAIL!

使用列表列或嵌套 data.frame 测试 tibbles 的相等性

Testing equality of tibbles with list-columns or nested data.frame

equality

r

testthat