使用R的bench包测试tidytable时出现结果不匹配错误

我尝试把tidytable包加入性能测试的基准测试中，在bench::mark里添加tidytable相关测试项后，弹出错误：

Error: Each result must equal the first result:
t_tidytable does not equal t_dplyr

对应的R代码如下：

library(dtplyr)
library(tidyverse)
library(profvis)
library(bench)
library(vctrs)
#> 
#> Attaching package: 'vctrs'
#> The following object is masked from 'package:dplyr':
#> 
#>     data_frame
#> The following object is masked from 'package:tibble':
#> 
#>     data_frame
library(data.table)
#> 
#> Attaching package: 'data.table'
#> The following objects are masked from 'package:lubridate':
#> 
#>     hour, isoweek, mday, minute, month, quarter, second, wday, week,
#>     yday, year
#> The following objects are masked from 'package:dplyr':
#> 
#>     between, first, last
#> The following object is masked from 'package:purrr':
#> 
#>     transpose

mtcars_tbl = tibble::as_tibble(mtcars, rownames = "make_model")

res =
  bench::mark(
    t_dplyr = dplyr::filter(mtcars_tbl, hp > 100),
    t_vctr = vec_slice(mtcars_tbl, mtcars_tbl$hp > 100),
    t_datatable = mtcars_tbl[mtcars_tbl$hp > 100, ]
  ) %>%
  select(expression, median)

res =
  bench::mark(
    t_tidytable = tidytable::filter(mtcars_tbl, hp > 100),
    t_dplyr = dplyr::filter(mtcars_tbl, hp > 100),
    t_vctr = vec_slice(mtcars_tbl, mtcars_tbl$hp > 100),
    t_datatable = mtcars_tbl[mtcars_tbl$hp > 100, ]
  ) %>%
  select(expression, median)
#> Error: Each result must equal the first result:
#> `t_tidytable` does not equal `t_dplyr`

Created on 2023-04-29 with reprex v2.0.2

这个错误本质是bench::mark默认会强制校验所有测试项的输出结果和第一个结果完全一致，但tidytable返回的是tidytable类对象，dplyr返回的是tibble类对象，二者的对象类型不同，导致校验失败。

有两种简单的解决途径：

1. 跳过一致性检查 #

直接在bench::mark里加check = FALSE参数，关闭结果相等性校验：

res =
  bench::mark(
    t_tidytable = tidytable::filter(mtcars_tbl, hp > 100),
    t_dplyr = dplyr::filter(mtcars_tbl, hp > 100),
    t_vctr = vec_slice(mtcars_tbl, mtcars_tbl$hp > 100),
    t_datatable = mtcars_tbl[mtcars_tbl$hp > 100, ],
    check = FALSE  # 新增参数
  ) %>%
  select(expression, median)

2. 统一输出对象类型 #

把不同方法返回的对象转换成同一种类型，比如把tidytable的结果转成tibble，或者反过来：

• 将tidytable结果转为tibble：

res =
  bench::mark(
    t_tidytable = as_tibble(tidytable::filter(mtcars_tbl, hp > 100)),
    t_dplyr = dplyr::filter(mtcars_tbl, hp > 100),
    t_vctr = vec_slice(mtcars_tbl, mtcars_tbl$hp > 100),
    t_datatable = mtcars_tbl[mtcars_tbl$hp > 100, ]
  ) %>%
  select(expression, median)

• 或者将所有结果转为tidytable：

res =
  bench::mark(
    t_tidytable = tidytable::filter(mtcars_tbl, hp > 100),
    t_dplyr = as_tidytable(dplyr::filter(mtcars_tbl, hp > 100)),
    t_vctr = as_tidytable(vec_slice(mtcars_tbl, mtcars_tbl$hp > 100)),
    t_datatable = as_tidytable(mtcars_tbl[mtcars_tbl$hp > 100, ])
  ) %>%
  select(expression, median)

额外验证 #

如果担心数据内容不一致，可以用all.equal()手动检查，忽略对象属性差异只看数据：

all.equal(
  tidytable::filter(mtcars_tbl, hp > 100),
  dplyr::filter(mtcars_tbl, hp > 100),
  check.attributes = FALSE
)

内容的提问来源于stack exchange，提问作者abalter