UTF+8 issue with rtf function, "N=" duplicated

[jr1234567]

Hello Many thanks for this great package I noticed that the rtf() function seems to add an extra "N=" string in table header of generated rtf file, as exemplified in the following example:

a <-   tangram('drug~bili["%4.03f"]+albumin+stage::Categorical[1]+protime+sex["%4.06f"]+age+spiders[1]', 
         data=pbc,
         pformat = 5,
         style="lancet",
         caption = "Table Lancet Style",
         relsize=-2,
         capture_units=FALSE,
         footnote = ""
 ) %>%  del_col(3)

  rtf(  a, file="file.rtf")

[spgarbet]

rtf gets the least love. This should fix it. Let me know if that works. Also, I noticed the lancet style has issues with the centered decimal. Do you see this on your end?

[jr1234567]

Many thanks. The duplicated "N=" is resolved. Indeed there are other issues: -With the lancet style the duplicated decimal gives a weird "Â" -With the Hmisc style, the "/" is missing, whereas the "*" are duplicated -With the nejm style, the dash is replaced by "0—"

[spgarbet]

These are all UTF+8 character handling. I need to look into what's required for this in RTF. The RTF spec was designed before unicode. I changed the title. Hopefully there's just some escaping required.

[spgarbet]

Found this: https://stackoverflow.com/questions/66275158/specify-utf-8-character-encoding-in-rtf-the-text-in-utf-8-format-is-correctly

[spgarbet]

https://www.oreilly.com/library/view/rtf-pocket-guide/9781449302047/ch01.html#unicode_in_rtf

I experimented with this. I can write a transformer I think.

There's also some Markdown not handled properly after the refactor. In particular, bold.

[spgarbet]

I've realigned this with the recent refactor. Should be able to move onto UTF+8 handling next.

[spgarbet]

Try this version.

[jr1234567]

• OK for the nejm style which appears to be rendered correctly
• With the lancet style : not OK: the centered decimal is rendered as ·
• With the hmisc style: not OK: the * are now missing

[spgarbet]

The hmisc "*" missing is intended. It should be bold now.

I'll check on the lancet. That was working on my side. The centered decimal is in the lower code space of UTF-8 which of course requires different handling.

[spgarbet]

I've duplicated the · on a different machine. Trying to solve.

[spgarbet]

It works for me now with all three cases. The only thing I see odd is the left column in 'nejm' is centered instead of left justified.

[jr1234567]

new tests:

• hmisc style : * and / are not rendered.
• lancet style : everything seems ok
• nejm style : everything seems ok

[spgarbet]

Why do you want the "*" rendered? It's a bold directive. I switched it to be consistent with the other styles and made it a bold.

[jr1234567]

Ah OK ! I misunderstood that,. I only mentioned them for consistency reason as that they do appear as such when viewing a table in Rstudio This is indeed way better not to have them
The / is useful though for numerator / denominator

[spgarbet]

I think this leaves rtf in good shape now.

[jr1234567]

Everything is beautiful Many thanks for this great job