Headerless tables not formatted correctly

[fkohrt]

Pandoc's simple tables may or may not contain a header¹. Headerless tables are, however, not formatted correctly by rmarkdown. The following document...

---
title: "Beheaded tables"
output: html_document
---

  Right     Left     Center     Default
-------     ------ ----------   -------
     12     12        12             12
    123     123       123           123
      1     1          1              1

Table: Simple table with header


-------     ------ ----------   -------
     12     12        12             12
    123     123       123           123
      1     1          1              1
-------     ------ ----------   -------

Table: Simple table without header

...renders to:

I ran into this while attempting to write a table with a header column – for which a syntax is underway, but not available right now – and hence without a header row. My attempt to overcome rmarkdown's limitation is as follows (based on this solution by Jan):

---
title: "Patched up tables"
output: html_document
---

```{r, echo=FALSE}
table_header <- c("Title", "Author", "Version", "Last edit", "License", "Address", "Contact")
table_body <- c("Headerless tables not formatted correctly",
                "fkohrt", "1.0.0", "2020-09-08", "CC0 1.0",
                "https://github.com/rstudio/rmarkdown/issues/1893",
                "https://github.com/fkohrt")

table_frame <- data.frame(table_body)
row.names(table_frame) <- table_header
table <- kableExtra::kable_styling(kableExtra::column_spec(knitr::kable(table_frame, format="html", caption="Not so simple table without header row"), 1, bold=TRUE), bootstrap_options = c("condensed"))
gsub("<thead>.*</thead>", "", table)
```

...which looks like this:

But I believe there's no reason to limit R Markdown to ... capitated tables.

xfun::session_info('rmarkdown')

R version 3.6.3 (2020-02-29)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Ubuntu 18.04.4 LTS, RStudio 1.2.5001

Locale:
  LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C               LC_TIME=en_US.UTF-8        LC_COLLATE=en_US.UTF-8    
  LC_MONETARY=en_US.UTF-8    LC_MESSAGES=en_US.UTF-8    LC_PAPER=en_US.UTF-8       LC_NAME=C                 
  LC_ADDRESS=C               LC_TELEPHONE=C             LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C       

Package version:
  base64enc_0.1.3 digest_0.6.25   evaluate_0.14   glue_1.4.2      graphics_3.6.3  grDevices_3.6.3 highr_0.8      
  htmltools_0.5.0 jsonlite_1.7.0  knitr_1.29      magrittr_1.5    markdown_1.1    methods_3.6.3   mime_0.9       
  rlang_0.4.7     rmarkdown_2.3.3 stats_3.6.3     stringi_1.4.6   stringr_1.4.0   tinytex_0.25    tools_3.6.3    
  utils_3.6.3     xfun_0.16       yaml_2.2.1     

Pandoc version: 2.10.1

¹ as can multiline tables and grid tables, but not pipe tables

---

By filing an issue to this repo, I promise that

• I have fully read the issue guide at https://yihui.org/issue/.
• I have provided the necessary information about my issue.
  • If I'm asking a question, I have already asked it on Stack Overflow or RStudio Community, waited for at least 24 hours, and included a link to my question there.
  • If I'm filing a bug report, I have included a minimal, self-contained, and reproducible example, and have also included xfun::session_info('rmarkdown'). I have upgraded all my packages to their latest versions (e.g., R, RStudio, and R packages), and also tried the development version: remotes::install_github('rstudio/rmarkdown').
  • If I have posted the same issue elsewhere, I have also mentioned it in this issue.
• I have learned the Github Markdown syntax, and formatted my issue correctly.

I understand that my issue may be closed if I don't fulfill my promises.

[cderv]

Thanks for the feedback ! I need some precision because I am not sure of what is the expected output. I tested with pandoc only and I got the same as with Rmarkdown

I put this in test.md

-------     ------ ----------   -------
     12     12        12             12
    123     123       123           123
      1     1          1              1
-------     ------ ----------   -------

Table: Simple table without header

And call

pandoc -f markdown -t html4 -o test.html test.md

I got this in test.html

Converting from rmarkdown

---
output: html_document
---

-------     ------ ----------   -------
     12     12        12             12
    123     123       123           123
      1     1          1              1
-------     ------ ----------   -------

Table: Simple table without header

gives me the same (with some styling minor difference)

Are you expecting something else ?

I am trying to see where Rmarkdown could interfere. The md content is untouched for pandoc to convert, and html_document used markdown as from format for pandoc which as simple_tables extension activated by default.

Thanks!

[fkohrt]

Yeah, my bad, I had not looked enough at the resulting HTML, which, in principle is fine indeed.

What I am expecting when converting the following document...

---
output: html_document
---

-------     ------ ----------   -------
     12     12        12             12
    123     123       123           123
      1     1          1              1
-------     ------ ----------   -------

Table: Simple table without header

...is this...

...instead of this...

The difference between them is the class table table-condensed that is being applied to the <table> in the first screenshot. It needs to be applied for headerless tables as well.

[cderv]

Thank you for the precision! I agree it looks better !

From my test and example with the command line, it seems pandoc itselfs does not output your expected table.
Do you agree on that ? Or did I miss something ? Do you get the correct formatting using pandoc (without rmarkdown) ?

[atusy]

table-condensed class is added by JavaScript below, which only targets at tables with headers.

rmarkdown/inst/rmd/h/default.html

Line 562 in c3b37f2

$$('tr.header').parent('thead').parent('table').addClass('table table-condensed');

[fkohrt]

Yeah, I believe it's not an issue with Pandoc, but with rmarkdown's HTML template.

From my test and example with the command line, it seems pandoc itselfs does not output your expected table.
Do you agree on that ?

So yeah, Pandoc does not output my expected table, but Pandoc does neither for tables with headers. The styling is added by rmarkdown, but only for tables with headers.

Do you get the correct formatting using pandoc (without rmarkdown) ?

Pandoc doesn't provide formatting (as in: CSS) at all by default.

[cderv]

I know understand the issue ! Sorry for all the question, it wasn't clear enough to me to know where to look.

I think @atusy pointed to the correct part - we will see how to support header less tables.

Thanks for the feedback @fkohrt !

[yihui]

@fkohrt You could add the classes by yourself via a js code chunk:

---
title: "Beheaded tables"
output: html_document
---

  Right     Left     Center     Default
-------     ------ ----------   -------
     12     12        12             12
    123     123       123           123
      1     1          1              1

Table: Simple table with header


-------     ------ ----------   -------
     12     12        12             12
    123     123       123           123
      1     1          1              1
-------     ------ ----------   -------

Table: Simple table without header

```{js, echo=FALSE}
$('table').addClass('table table-condensed'); 
```

We may be able to do this in the default HTML template, but I'm not sure if it could break anything, given that the template has existed for quite a few years.

[cderv]

I just opened a PR with a quick fix to discuss if it worth taking the risk that something break or not.
The current selector has been around this 2016 and rmarkdown 0.6 (7d29206) - so yeah quite some time !

[fkohrt]

I saw people struggle with this before, so I'm not sure of no action is the best bet in the long run. But you could wait until the dust settles with the AST change to tables, i.e. the readers and writers get implemented.

The immediate problem I had is remedied by the two workarounds discussed above.

[github-actions]

This old thread has been automatically locked. If you think you have found something related to this, please open a new issue by following the issue guide (https://yihui.org/issue/), and link to this old issue if necessary.