bug fix: truncate wide columns in print.data.table #7718 by venom1204 · Pull Request #7758 · Rdatatable/data.table

venom1204 · 2026-05-25T18:29:50Z

I incorporated the changes as proposed, below is a short summary of the changes

print.data.table.r - i updated the char.trunc to dafualt getOption("width") - 5L, and also added condition to apply for lists. improved safety for strings with missing width values.

Kindly review this when you have some time, and please let me know if you have any suggestions for improvement.
thank you.

codecov · 2026-05-25T18:36:17Z

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 99.04%. Comparing base (d4974e9) to head (1af6897).
⚠️ Report is 13 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #7758      +/-   ##
==========================================
- Coverage   99.04%   99.04%   -0.01%     
==========================================
  Files          87       87              
  Lines       17064    17063       -1     
==========================================
- Hits        16901    16900       -1     
  Misses        163      163

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

joshhwuu

thanks, left a few comments! in the future, you can tag me as a reviewer instead of pinging me in the description. also please try to summarize changes in the description instead of referring to the issue so others can get context quicker 🙂

joshhwuu · 2026-05-25T18:39:47Z

 # Current implementation may have issues when dealing with strings that have combinations of full-width and half-width characters,
 # if this becomes a problem in the future, we could consider string traversal instead.
 char.trunc = function(x, trunc.char = getOption("datatable.prettyprint.char")) {
+  if (is.null(trunc.char)) trunc.char = getOption("width") - 5L


just curious, why 5L? and this looks like a magic number, could we use a constant instead?

my major reason of using 5L was because toby sir mentioned this would options(datatable.prettyprint.char=getOption("width")-5) be possible? in the issue

would 5L still make sense if we had a large table and the row label prefix consists of many more characters? I would argue against a magic number or even reading the row label prefix if we can here

you are right the 5L buffer fails once row labels grow bigger on a narrow consoles.
to fix this we can either

increase the constant biffer instead of using 5L

or we can pass the width dyanamically for precise fit.

can you please provide a couple examples of when 5L works and does not work?
does is work for the examples I put in the original issue?

its working fine for your example ,i was accounting about the final width because the 5L bugffer doent not account all layout components (row labels,spacing,'...')
for ex

options(width=25, datatable.prettyprint.char=NULL) dt = data.table( x = rep("ABCDEFGHIJKLMNOPQRSTUVWXYZ", 1000000) ) print(dt, topn=1) >1000000: ABCDEFGHIJKLMNOPQRST...

if you count the truncated data only then its behaving correctly , however if teh fully rendered line is considered then its excedding the 25 limit.

thanks. options(width) is documented on ?options as a limit on the whole line (the maximum number of columns on a line used in printing vectors, matrices and arrays, and when filling by cat), so in this example the output should be 25 characters,

1000000: ABCDEFGHIJKLM...

can you guide me towards what could be the next step.

I’m not sure, I have not thought it through.

joshhwuu · 2026-05-25T18:47:06Z

+  nchar_chars = nchar(x, 'char', allowNA=TRUE)
  is_full_width = nchar_width > nchar_chars
-  idx = !is.na(x) & pmin(nchar_width, nchar_chars) > trunc.char
+  idx = !is.na(x) & !is.na(nchar_width) & pmin(nchar_width, nchar_chars) > trunc.char


could we use na.rm in pmin instead of a guard?

i tried using na.rm = TRUE in pmin but found it causes NA values to be incorrectly flagged for truncation which leads to NA becoming "NA..." (breaking test 2253.2) and invalid 'width' argument crashes in strtrim.

github-actions · 2026-05-25T18:54:26Z

HEAD=issue77188 slower P<0.001 for DT[by] max regression fixed in #7480

Generated via commit 1af6897

Download link for the artifact containing the test results: ↓ atime-results.zip

Task	Duration
R setup and installing dependencies	6 minutes and 22 seconds
Installing different package versions	12 minutes and 5 seconds
Running and plotting the test cases	5 minutes and 23 seconds

tdhock · 2026-05-28T13:04:25Z

+options(width=10, datatable.prettyprint.char=NULL)
+test(2374.1, capture.output(print(data.table(x="12345678901234567890"))), output="12345...")


please use test(options=) instead of options(…), here and below

Suggested change

options(width=10, datatable.prettyprint.char=NULL)

test(2374.1, capture.output(print(data.table(x="12345678901234567890"))), output="12345...")

test(2374.1, capture.output(print(data.table(x="12345678901234567890"))), output="12345...", options=list(width=10, datatable.prettyprint.char=NULL))

Co-authored-by: Toby Dylan Hocking <toby.dylan.hocking@usherbrooke.ca>

This reverts commit e2095c1.

* prohibit duplicate key columns * update test setup * adjust test setups * add Mikes found cases * use CJ(, sorted=FALSE) and sort afterwards * cache grpnames

* add setallocrow * fix copyAsPlain args * change allocrow(dt, n) n parameter to specify number of nrows and not number of additional rows * remove repetition in docs, note address change

* `between` supports Date/IDate with missing bounds * Restore NEWS formatting * rm unnecessary lines from tests.Rraw

* Data desegregation + minor formatting improvements + constness improvements --------- Co-authored-by: Benjamin Schwendinger <52290390+ben-schwen@users.noreply.github.com>

update documentation to clarify zero-length RHS evaluations with/without by

* docs: add datatable.prettyprint.char to datatable-intro vignette Closes #7714 Added explanation and example of datatable.prettyprint.char option in the 'Note that' section, near the datatable.print.nrows bullet.

* don't save to var as suggested by Toby Co-authored-by: Toby Dylan Hocking <tdhock5@gmail.com> --------- Co-authored-by: Toby Dylan Hocking <tdhock5@gmail.com>

Since R-2.15, R automatically replaces dots with understores in R_init_<DLL name>(...), so renaming is no longer necessary. This also has the side effect of working around the behaviour change from R r90101 and producing correct source tarballs on Windows again: R CMD build cleans up data.table.dll, but misses data_table.dll. * Accomodate new DLL name in atime tests

venom1204 added 2 commits May 25, 2026 12:21

proper checks

112f675

news entry

1af6897

venom1204 requested a review from MichaelChirico as a code owner May 25, 2026 18:29

joshhwuu reviewed May 25, 2026

View reviewed changes

venom1204 requested a review from joshhwuu May 27, 2026 21:08

tdhock requested changes May 28, 2026

View reviewed changes

venom1204 mentioned this pull request Jun 1, 2026

main discussion page for GSOC-2026 (VINIT) #1 venom1204/discussion-GSOC-2026#1

Open

tdhock and others added 21 commits June 10, 2026 19:44

link datatableAPI.h, rm non-r (#7756)

2ff111e

Co-authored-by: Toby Dylan Hocking <toby.dylan.hocking@usherbrooke.ca>

revert PR 7730 until yihui/actions is fixed (#7761)

b4d2187

Revert "revert PR 7730 until yihui/actions is fixed (#7761)" (#7763)

5d9e6ac

This reverts commit e2095c1.

update documentation to reflect supported types for nafill (#7762)

5858d92

prohibit duplicate key columns (#7760)

9e06c51

* prohibit duplicate key columns * update test setup * adjust test setups * add Mikes found cases * use CJ(, sorted=FALSE) and sort afterwards * cache grpnames

add setallocrow (#7644)

a676c02

* add setallocrow * fix copyAsPlain args * change allocrow(dt, n) n parameter to specify number of nrows and not number of additional rows * remove repetition in docs, note address change

fix clang CI warnings (#7767)

d8aef4e

markdown fields in revdep_check_failure.yml (#7771)

bad8536

between supports Date/IDate with missing bounds (#7751)

c571e33

* `between` supports Date/IDate with missing bounds * Restore NEWS formatting * rm unnecessary lines from tests.Rraw

Declare variables at point of use in assign.c (#7303)

d756520

* Data desegregation + minor formatting improvements + constness improvements --------- Co-authored-by: Benjamin Schwendinger <52290390+ben-schwen@users.noreply.github.com>

Documented zero-length RHS behavior with by (#7768)

eaf0e75

update documentation to clarify zero-length RHS evaluations with/without by

docs: add datatable.prettyprint.char to datatable-intro vignette (#7765)

8e9a291

* docs: add datatable.prettyprint.char to datatable-intro vignette Closes #7714 Added explanation and example of datatable.prettyprint.char option in the 'Note that' section, near the datatable.print.nrows bullet.

remove DT overriding to not break vignette (#7773)

3200f30

* don't save to var as suggested by Toby Co-authored-by: Toby Dylan Hocking <tdhock5@gmail.com> --------- Co-authored-by: Toby Dylan Hocking <tdhock5@gmail.com>

Fix indent in vignette code block (#7775)

1f12cf1

Fix name spellings and add ORCID. (#7787)

39d92ee

proper checks

80c3c64

news entry

2d3cab7

random

3e190ac

just fixed

0048e66

..

5b1f973

..

a872aa4

venom1204 requested review from Anirban166 and HughParsonage as code owners June 10, 2026 19:45

venom1204 closed this Jun 10, 2026

venom1204 deleted the issue77188 branch June 10, 2026 19:46

venom1204 restored the issue77188 branch June 10, 2026 19:48

venom1204 deleted the issue77188 branch June 10, 2026 19:51

		options(width=10, datatable.prettyprint.char=NULL)
		test(2374.1, capture.output(print(data.table(x="12345678901234567890"))), output="12345...")

Conversation

venom1204 commented May 25, 2026

Uh oh!

codecov Bot commented May 25, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

joshhwuu left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

venom1204 May 26, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

github-actions Bot commented May 25, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

10 participants

codecov Bot commented May 25, 2026 •

edited

Loading

venom1204 May 26, 2026 •

edited

Loading