docx+native_numbering is ignored and subfigure not properly exportes to Docx

[aminevsaziz]

Is it normal that quarto render keep ignoring docx+native_numbering when exporting to Docx format. Pamdoc on the other hand work fine with native numbering,why is that?, if yes what the solution to have native numbering in docx?
Also sub figures are mess when exported to Docx? Am I missing something or subfigure syntax only work for html export?

Here are an example with the screenshot of the output

---
format:
  docx:
    #standalone: true
    variant: +native_numbering
crossref: true
---

::: {#fig-elephants layout-ncol=2}

![Surus](surus.png){#fig-surus}

![Hanno](hanno.png){#fig-hanno}

Famous Elephants
:::

[cderv]

I can't say for sure why +native_numbering does not work specifically, but it is probably related to Quarto own crossreference system https://quarto.org/docs/authoring/cross-references.html

You are using #fig- label, so you are using Quarto own system that takes charge of numbering - This means the own pandoc system may not work and be adapted to the quarto system.

Can you explain your expectation ? By using Quarto's cross references system, you should not expect Pandoc's own system to work. And this is why I am interested to understand what is missing and not working for you.

Also, know that for Quarto Cross Ref to work in Docx, some layout are used that may not be what you expect

• Both figure and table captions are being converted to actual table structures within Word. This results in the captions being displayed as tables, not text, completely breaking the document layout and structure. #11756

This is a known limitations for now.

[aminevsaziz]

I can't say for sure why +native_numbering does not work specifically, but it is probably related to Quarto own crossreference system https://quarto.org/docs/authoring/cross-references.html

You are using #fig- label, so you are using Quarto own system that takes charge of numbering - This means the own pandoc system may not work and be adapted to the quarto system.

Can you explain your expectation ? By using Quarto's cross references system, you should not expect Pandoc's own system to work. And this is why I am interested to understand what is missing and not working for you.

Also, know that for Quarto Cross Ref to work in Docx, some layout are used that may not be what you expect

• Both figure and table captions are being converted to actual table structures within Word. This results in the captions being displayed as tables, not text, completely breaking the document layout and structure. #11756

This is a known limitations for now.

Well if i keep the same document and use quarto pandoc to render the document I get Figure 1 and Figureand number 1 and 2can be updated in MS word as field and the fenced div of subfigure layout is ignored completely. However, when I use only quarto render the subfigure fenced div is acknowledged but figures are not well formated (weird table around figures) and caption and numbering are just pure text and can not be updated as field.

What I expect was the following:

• having subfigure layout (e.g 2 columns) that work in docx similar to html.
• the produced float (figure, table...etc) numbering is not plain text but rather fields that MS Word recognize.

[cderv]

Well if i keep the same document and use quarto pandoc to render the document

To be clear, when you do that, you are not using quarto render and so no Quarto specific feature will work. In that case you are using pandoc conversion features.

when I use only quarto render the subfigure fenced div is acknowledged but figures are not well formated (weird table around figures)

This is exactly #11756 - the table around figures is a trick to be able to make the cross ref feature work at the time it was implemented. We need to improve this now.

the produced float (figure, table...etc) numbering is not plain text but rather fields that MS Word recognize.

This is the hard part. It requires to have cross ref feature use this MS Word feature through Pandoc, and as of now this was not possible. In age of AI, maybe this can be retackle. It also depends what openxml and MS Words allows.

Maybe we can plug in +native-numbering with cross ref system, but it is a rather complex feature that needs to work the same on various format, and for now the way it was made available to DOCX is through the current output.

So you should follow #11756 which is the feature request associated with this

[aminevsaziz]

Interesting since 2024 or maybe earlier and this issue is not solved 😢
So it seems that quarto is useful when exporting for html/pdf (maybe other formats). Otherwise exporting to Docx seems to work on paper. I hope quarto 2 or maybe next release solve this issue 🤞

[cderv]

Otherwise exporting to Docx seems to work on paper

It works but not to the extent to your expectations. We are doing our best with what time allows, with a focus on q2 in recents month that will support producing docx too.

You're free to use Pandoc directly if what they offers for docx fits your need, though format: docx still produce decent documents that will have same features as other html or pdf formats, in the limits of what is possible and offered by the format.

Issue is still open as you saw, so we plan to find a solution at some point for this specific limitation, but docx is a supported format for sure, and not just on paper.

Anyhow, we've heard the feedback and the interesting in fixing this.