teitodocx​​ does not​ ​correctly ​handle ​​rows attribute of​ ​a table​ ​cell

[RayCHOU]

I have a TEI table​ shown like as following: ​

<table>
  <row>
     <cell>A</cell>
     <cell rows="2">B</cell>
  </row>
  <row>
     <cell>C</cell>
  </row>
</table>

I ​try to ​use teitodocx​​ to convert this ​TEI file to docx​ format​​, and I expect to get result like:

But result is:

any suggestions?​

[jamescummings]

Tested and confirmed that the result is as you say. (Well, in my case the C cell stretches the full width....) I tried a variety of likely markup to see if that triggered a conversion spanning rows and it didn't, so I looked at the XSL.

If someone is trying to fix this the table-handling in teitodocx.xsl is at https://github.com/TEIC/Stylesheets/blob/dev/docx/to/teitodocx.xsl#L1237 and the templates for rows and cells at https://github.com/TEIC/Stylesheets/blob/dev/docx/to/teitodocx.xsl#L1399

I think it has logic built in for @cols but doesn't do anything with @row. There is, in calculating the table borders some handling of html:colGroup and cals: tables but again I think this is handling column spanning and grouping rather than row spanning. Fixing it is non-trivial I think because it will conflict with the way we're doing the column spanning.

@RayCHOU: I recognise confirming that it doesn't work (and seeing that the XSLT doesn't handle @rows) isn't much help. If you need to get lots of TEI tables as marked up in your example I would probably suggest transforming them to two nested tables. If you don't have many of them, then I'd be tempted to suggest changing them by hand after the fact.

[Cutuchiqueno]

Does anyone had some ideas on this meanwhile or are there other approaches than nest two tables (which seams too dirty for me)?

In a single-source publishing project there will be a lot of 300 pages Word documents with around 50-80 tables each, many of them using a complicated table structure. Fixing it by hand will be a lot of work and it will be open ended.

I would be willing to have a look at the XSLT though at the moment it is out of my technical capacity.