user specified width does not override getOption("width")

[jtlandis]

Hello Dev team,

I am unsure if the is intentional or not. I found a behavior in which extra columns that do not fit along getOption("width") are wrapped onto the new line. Is the expected usage for the width argument in tibble::print() to be left NULL now?

[krlmlr]

Thanks. Can you please share your example?

[jtlandis]

Sure! I didnt have time to write a good regex earlier, consider the following (only really applicable outside of Rstudio, so in an R session started via a terminal)

library(tibble)
data <- lapply(1:20, \(x) sample(1:200, 5)) |> 
  setNames(nm = sprintf("column_%i", 1:20)) |> 
  as_tibble()
getOption("width")
#> [1] 80

printing wraps with specified width is > “width” option

print(data, width = 100)
#> # A tibble: 5 × 20
#>   column_1 column_2 column_3 column_4 column_5 column_6 column_7 column_8
#>      <int>    <int>    <int>    <int>    <int>    <int>    <int>    <int>
#> 1        9       37        8      195       16      179      122       84
#> 2       69      193        1       34       17       91       45        4
#> 3      113       23       34       11      158      116       10       82
#> 4      172      102       77      149      159      184       80      126
#> 5       61       96      104      152      120      101        2       78
#>   column_9 column_10 column_11 column_12 column_13 column_14 column_15 column_16
#>      <int>     <int>     <int>     <int>     <int>     <int>     <int>     <int>
#> 1       76       116        49       103       189        58       195        38
#> 2      168         4        24        49        76       150       163        47
#> 3       66       155        43        42       103        21       112       122
#> 4      101        83        93        34       151       135        45        59
#> 5      124       190       125       165       112       184       129       168
#> # ℹ 4 more variables: column_17 <int>, column_18 <int>, column_19 <int>, column_20 <int>
options(width = 100)
print(data)
#> # A tibble: 5 × 20
#>   column_1 column_2 column_3 column_4 column_5 column_6 column_7 column_8 column_9 column_10
#>      <int>    <int>    <int>    <int>    <int>    <int>    <int>    <int>    <int>     <int>
#> 1        9       37        8      195       16      179      122       84       76       116
#> 2       69      193        1       34       17       91       45        4      168         4
#> 3      113       23       34       11      158      116       10       82       66       155
#> 4      172      102       77      149      159      184       80      126      101        83
#> 5       61       96      104      152      120      101        2       78      124       190
#> # ℹ 10 more variables: column_11 <int>, column_12 <int>, column_13 <int>, column_14 <int>,
#> #   column_15 <int>, column_16 <int>, column_17 <int>, column_18 <int>, column_19 <int>,
#> #   column_20 <int>

data wraps again

print(data, width = 120)
#> # A tibble: 5 × 20
#>   column_1 column_2 column_3 column_4 column_5 column_6 column_7 column_8 column_9 column_10
#>      <int>    <int>    <int>    <int>    <int>    <int>    <int>    <int>    <int>     <int>
#> 1        9       37        8      195       16      179      122       84       76       116
#> 2       69      193        1       34       17       91       45        4      168         4
#> 3      113       23       34       11      158      116       10       82       66       155
#> 4      172      102       77      149      159      184       80      126      101        83
#> 5       61       96      104      152      120      101        2       78      124       190
#>   column_11 column_12 column_13 column_14 column_15 column_16 column_17 column_18 column_19
#>       <int>     <int>     <int>     <int>     <int>     <int>     <int>     <int>     <int>
#> 1        49       103       189        58       195        38       109       131       117
#> 2        24        49        76       150       163        47       165       198       194
#> 3        43        42       103        21       112       122       175       167        71
#> 4        93        34       151       135        45        59       110       101       183
#> 5       125       165       112       184       129       168        98       122        37
#> # ℹ 1 more variable: column_20 <int>

if this is the intended behavior, should it be documented?
or maybe it is already documented in pillar/tibble and I have
missed it?

^{Created on 2024-09-14 with reprex v2.1.1}

[krlmlr]

Thanks. Are any of the options in https://pillar.r-lib.org/reference/pillar_options.html set on your system?

[jtlandis]

I was not previously setting these options, but I was able to get the desired effect with options(width = 120L), and I see that it is documented there. Thanks! Do you think it would be worth adding some documentation to ?tibble::print.tbl, since the width argument implies that the options are only consulted when width = NULL

Width of text output to generate. This defaults to NULL, which means use the width option.