Alternative method to display numeric values that were truncated for display

[lorenzwalthert]

Essentially migrating tidyverse/tibble#362 (comment).

with pillar 1.2.1 and tibble 1.4.2 I get this

library(tibble)
library(purrr)
tibble(
  integer = sample(100),
  numeric = integer + runif(100, 0, 10e-4),
  long_string = rerun(100, paste0(sample(rep(letters, 10)), collapse = "")) %>% flatten_chr()
)
#> # A tibble: 100 x 3
#>    integer numeric long_string                                            
#>      <int>   <dbl> <chr>                                                  
#>  1      51    51.0 zwsbmmgoegfginnikiyidrqcvkbivgkqtqpeblkezcntutkrmzchcm…
#>  2      37    37.0 qovlyzbtbyrfjnrxvitqdvjefyeskwfqdkkxjgirceepdnyzhglgas…
#>  3      33    33.0 lkfgpmzekxvtdlytkzhqkxunuhccfspefajweoiaeficvzgqqpvevf…
#>  4      69    69.0 fgihsvbyqcxbmynuxuywliptbabzomrbnvgbdelvsqmmscmwiknskg…
#>  5      84    84.0 zlrfncxjqvtllvmmzddslyeoiwtswyaexnxkfmwebbxmpmltnstfhe…
#>  6      60    60.0 lunkdilxuipgkvnixxumjsnqilcqskqyyxeyxpmpzavroxjqauwcga…
#>  7      77    77.0 wklbnoexoyimgdwyywzoosnupqejygsahxxfmugwynhnqfallvrqvc…
#>  8      74    74.0 rgrtqshfbxomhhilvurhmdqgwuveeldugulmtueybznfvrcpqvwiia…
#>  9      91    91.0 iswngaunersfosinwepihmznjwqxfeitmdgkisthugggxmggovsvyc…
#> 10      21    21.0 akadobtgrfnhjgbndrgglfalglvzkifibxkniqzqizexyfyfcwiicw…
#> # ... with 90 more rows

Created on 2018-03-02 by the reprex package (v0.2.0).

I think it's hard to guess that 96.0 does not represent the same value as 96, at least for me. Hence, I propose to represent numbers that were truncated for display (more) similar to truncated text: With three trailing dots after the decimal dot (that's at least what it looks like on my mac).

As pointed out by @krlmlr in tidyverse/tibble#362 (comment), I could increase the pillar option to show more digits, but I guess my point is not that I want to know the exact number, but rather I think it would be beneficial to see at first glance whether a number is truncated for display or not, in an intuitive way. So the above would become

library(tibble)
library(purrr)
tibble(
  integer = sample(100),
  numeric = integer + runif(100, 0, 10e-4),
  long_string = rerun(100, paste0(sample(rep(letters, 10)), collapse = "")) %>% flatten_chr()
)
#> # A tibble: 100 x 3
#>    integer numeric long_string                                            
#>      <int>   <dbl> <chr>                                                  
#>  1      51    51.… zwsbmmgoegfginnikiyidrqcvkbivgkqtqpeblkezcntutkrmzchcm…
#>  2      37    37.… qovlyzbtbyrfjnrxvitqdvjefyeskwfqdkkxjgirceepdnyzhglgas…
#>  3      33    33.… lkfgpmzekxvtdlytkzhqkxunuhccfspefajweoiaeficvzgqqpvevf…
#>  4      69    69.… fgihsvbyqcxbmynuxuywliptbabzomrbnvgbdelvsqmmscmwiknskg…
#>  5      84    84.… zlrfncxjqvtllvmmzddslyeoiwtswyaexnxkfmwebbxmpmltnstfhe…
#>  6      60    60.… lunkdilxuipgkvnixxumjsnqilcqskqyyxeyxpmpzavroxjqauwcga…
#>  7      77    77.… wklbnoexoyimgdwyywzoosnupqejygsahxxfmugwynhnqfallvrqvc…
#>  8      74    74.… rgrtqshfbxomhhilvurhmdqgwuveeldugulmtueybznfvrcpqvwiia…
#>  9      91    91.… iswngaunersfosinwepihmznjwqxfeitmdgkisthugggxmggovsvyc…
#> 10      21    21.… akadobtgrfnhjgbndrgglfalglvzkifibxkniqzqizexyfyfcwiicw…
#> # ... with 90 more rows

Created on 2018-03-02 by the reprex package (v0.2.0).

[krlmlr]

The numbers in the numeric column are truncated not for display, but because by default only 3 significant figures are requested. This is what is shown in this column. The .0 suggests that there's more, whole numbers don't have the 0 after the decimal dot.

We see more if we ask for 6 significant figures:

library(tibble)
library(purrr)
options(pillar.sigfig = 6)
tibble(
  integer = sample(100),
  numeric = integer + runif(100, 0, 10e-4),
  long_string = rerun(100, paste0(sample(rep(letters, 10)), collapse = "")) %>% flatten_chr()
)
#> # A tibble: 100 x 3
#>    integer numeric long_string                                            
#>      <int>   <dbl> <chr>                                                  
#>  1      97 97.0001 bggdngfamupzwnoqtgnpeqxgepycbzoojruswgoucxowflljkhonep…
#>  2      13 13.0009 mzxwfaftnmugqiaukxjkwxzqmorukcufbsxaklttwsalglxpsuiijw…
#>  3      12 12.0008 zomszzwwrvajedxgaqhomgivvgdltlyvhvhcdckpfxxljeunfjupcc…
#>  4      83 83.0004 niekxtluwrkifjgiuovvgwzudsugibqkfmlkdcuzdjygpedlobbhym…
#>  5      58 58.0001 zzfieszqpotnneexxngltoemlqptkqbpmlptwpnegctcvdfcvtmonr…
#>  6      93 93.0006 qzdafvochukcfpazpsbgybtgncdxchvxatuzqivrrwhefogzimspuq…
#>  7      16 16.0005 rbfhrijvbskyamftsvwgduopeqevmivomepeqddjgrwldirnbmipwk…
#>  8      18 18.0003 iyvvaxwugorelcghunzcbqlrtewissyfpsxkedrxczjkzxvavqvhtn…
#>  9      38 38.0002 przmtjfypppokwmgbcxbuqxmqigwkllyuewbiqvialnyneesysupoq…
#> 10      21 21.0004 ztmqatqjxswfyqgwdpwwxwjnsunmpqizcofrmxbxcczjirysruvuod…
#> # ... with 90 more rows

Created on 2018-03-03 by the reprex package (v0.2.0).

Closing, please open a new issue in tibble if you still think the behavior is wrong even with more significant digits shown.

[github-actions]

This old thread has been automatically locked. If you think you have found something related to this, please open a new issue and link to this old issue if necessary.