r - How to use mathematical notation or engineering notation in certain columns of a kableExtra table?

Question

One can apply docxtools::format_engr() to a table to switch from computer notation to engineering notation.

(Source: docxtools)

Applying docxtools::format_engr() to a df changes all columns to engineering notation.

1. How to apply the function only to some columns of a df?
2. Is there an equivalent function for mathematical notation?

For the toy example below, the idea would be to get the p_Pa column in engineering/mathematical notation, while all other columns remain the same.

MWE

Table as desired only missing the p_Pa column in engineering notation

library(docxtools)
density2 <- density
density2$density <- density2$density*1000
kable(density2, digits = 2)

Apply formatting: "all" columns change to engineering notation

# Apply formatting:
density_engr <- density2 %>%
  docxtools::format_engr()
kable(density_engr)

Passing sigdig() to docxtools::format_engr() (e.g., sth. like "sigdig = c(0,0,2,3,2,3)") to define how to format which column does not help here. Engineering notation is still applied to ALL columns, only with differing numbers of digits.

The idea is to include the df as kableExtra table in an R markdown report compiled with knitr both to PDF and HTML.

Answer 1

You can create a dataframe that contains only the columns you want formatted, format that, and then assign it into place in the original. For example, starting with your density2:

density_engr <- density2        # no formatting yet
do_format <- c("T_K", "p_Pa")   # which columns to format?
density_engr[do_format] <- docxtools::format_engr(density_engr[do_format])

There are other ways to do the indexing besides density_engr[do_format], but be careful: you want to make sure you get a dataframe as the result, you don't want to extract a column.