Inconsistent table numbering - HTML vs. PDF #13063
Replies: 2 comments 3 replies
-
I don't know what your understanding of The LaTeX package As a result, it is hard for Quarto documents to work well with
In general, if you're not sure it's a bug, we ask you to open a discussion first. I'll move this to a discussion and we can take it from there. |
Beta Was this translation helpful? Give feedback.
-
|
Let's look at the first code cell in isolation. 13063.qmdkeep-md output: 13063.pdf.mdThe issue, as I suspected, is that In this case, Quarto does the right thing (not emitting a surrounding But if you use 13063-b.qmdkeep-md output: 13063-b.pdf.mdNote the In simple settings, Quarto actually attempts to detect |
Beta Was this translation helpful? Give feedback.
-
|
Thank you for the break down on all of that. I am nowhere near as familiar with LaTeX as I am R, so that's where my understanding tends to fall apart. However, I get the general gist of you have laid out. I had a sneaking suspicion it was related to I appreciate the input - I'll have to scratch my head on the best way to update our report code to ensure those tables still format across pages properly w/o incrementing the table counter. Thanks! |
Beta Was this translation helpful? Give feedback.
-
|
@cscheid There is indeed an issue with flextable environment, but the focus on this is probably adding another one. The OP example is also using Try this ---
title: "PDF Table Numbering Example"
format:
pdf:
keep-tex: true
keep-md: true
---
```{r}
#| include: false
library(dplyr)
library(knitr)
iris_summary <- iris |> reframe(across(where(is.numeric), mean), .by = Species)
iris_unlabeled <- iris |> filter(Species == "setosa") %>% head(4)
```
```{r}
#| label: tbl-kable1
#| tbl-cap: "This is a labeled kable"
#| echo: false
kable(iris_summary)
```
```{r}
#| echo: false
#| warning: false
#| message: false
kable(iris_unlabeled)
```
```{r}
#| label: tbl-kable2
#| tbl-cap: "This is the second labeled kable"
#| echo: false
kable(iris_summary)
```No flextable and this leads to this keep-md output ---
title: "PDF Table Numbering Example"
format:
pdf:
keep-tex: true
keep-md: true
---
::: {#tbl-kable1 .cell tbl-cap='This is a labeled kable'}
::: {.cell-output-display}
|Species | Sepal.Length| Sepal.Width| Petal.Length| Petal.Width|
|:----------|------------:|-----------:|------------:|-----------:|
|setosa | 5.006| 3.428| 1.462| 0.246|
|versicolor | 5.936| 2.770| 4.260| 1.326|
|virginica | 6.588| 2.974| 5.552| 2.026|
:::
:::
::: {.cell}
::: {.cell-output-display}
| Sepal.Length| Sepal.Width| Petal.Length| Petal.Width|Species |
|------------:|-----------:|------------:|-----------:|:-------|
| 5.1| 3.5| 1.4| 0.2|setosa |
| 4.9| 3.0| 1.4| 0.2|setosa |
| 4.7| 3.2| 1.3| 0.2|setosa |
| 4.6| 3.1| 1.5| 0.2|setosa |
:::
:::
::: {#tbl-kable2 .cell tbl-cap='This is the second labeled kable'}
::: {.cell-output-display}
|Species | Sepal.Length| Sepal.Width| Petal.Length| Petal.Width|
|:----------|------------:|-----------:|------------:|-----------:|
|setosa | 5.006| 3.428| 1.462| 0.246|
|versicolor | 5.936| 2.770| 4.260| 1.326|
|virginica | 6.588| 2.974| 5.552| 2.026|
:::
:::
No 
Isn't this a problem? I would expect the second table to not increment the counter. The .tex file has longtable env in the output despited being markdown in the source, so I believe this Pandoc writing longtable environment here. Can you reproduce this too? I want to open an issue about this and look into it. Then I'll open an issue with flextable if still present. I remember they did try to solve this (davidgohel/flextable#582), but I don't know what they did. I'll work with David on this. |
Beta Was this translation helpful? Give feedback.
-
Yes, I can. We should be fixing these tables and aren't. This is a bug and I don't know why our test suite isn't catching it. Very annoying. Even this is failing: |
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
-
I have:
Bug description
I am using Quarto with R to generate a large annual report based on long-term monitoring data. The report was drafted in Quarto beginning in 2022, and has been used ever since with minor script updates and tweaks to account for changes in wording, clarity, function deprecation, etc.
Last week I ran this year's report, and a colleague noted that a series of tables were misnumbered. That section's R script includes a labeled table followed by a series of unlabeled tables before another labeled table. This was done to prevent any one table in the first sequence from being split across pages (i.e., explicitly formatting two tables per page). So, the labeled tables jump from Table 2 to Table 8 (instead of Table 2 and then Table 3).
I reviewed previous reports (e.g., 2024, 2023), and those unlabeled tables were present with no breaks in the labeled-table number sequence.
I tried downgrading to past versions of
officerandflextable, as well as testing with previous, current, and pre-release versions ofQuarto. The behavior remains the same - unnumbered tables are increasing the table counter for labeled table for PDF but not for HTML. I updated Pandoc and Tinytex as well.I read some discussions/issues on various messages boards and repositories that suggested it may be specifically related to
flextableand something related to 'table' vs. 'longtable' float. However, that's where my knowledge base begins to drop off. Additionally, the behavior appears to persist when usingkable. Reviewing the intermediate .tex file {longtable} appears for allflextableandkabletables.I note old issues/discussions about this and recall a comment about a fix in v1.3 or 1.4. So, I'm not sure if this is a recent regression or I just have something crazy happening on my end that's beyond my scope of understanding. It is entirely possible that I don't understand how 'longtable' operates, and that I could simply call all 10 tables in that series at once and they will be formatted correctly across pages.
Apologies if this is not a but or not related to Quarto, and I appreciate any thoughts, input, or comments.
Steps to reproduce
Actual behavior
Unlabeled tables are increasing the table counter, breaking the number sequence for labeled tables - but only when rendering to PDF.
Expected behavior
Numbering sequence is unbroken for labeled tables, despite inclusion of unlabeled tables. This expected behavior was the case for past PDF documents.
Your environment
RStudio 2025.05.0+496 "Mariposa Orchid" Release (f0b76cc00df96fe7f0ee687d4bed0423bc3de1f8, 2025-05-04) for windows
Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) RStudio/2025.05.0+496 Chrome/132.0.6834.210 Electron/34.5.1 Safari/537.36, Quarto 1.8.16 (C:/Users/Michael.Schram/AppData/Local/Programs/Quarto/bin/quarto.exe)
Quarto check output
Beta Was this translation helpful? Give feedback.
All reactions