Chapter 11 Programming

11.1 Part dimensions

Programming with flextable can be made easier by using a few functions:

nrow_part() returns the number of lines in a part of flextable.
ncol_keys() returns the number of columns displayed.

For example, if you want to add an aggregation line below a sample and to format it differently from the sample rows :

library(flextable)
library(dplyr)
library(tibble)

dat_summary <- summarize(airquality, across(Ozone:Temp, ~ mean(.x, na.rm = TRUE))) %>% 
  add_column(label = "summary", .before = TRUE)
dat_summary

#     label    Ozone  Solar.R     Wind     Temp
# 1 summary 42.12931 185.9315 9.957516 77.88235

dat_sample <- head(airquality) %>% 
  add_column(label = "sample", .before = TRUE)
dat_sample

#    label Ozone Solar.R Wind Temp Month Day
# 1 sample    41     190  7.4   67     5   1
# 2 sample    36     118  8.0   72     5   2
# 3 sample    12     149 12.6   74     5   3
# 4 sample    18     313 11.5   62     5   4
# 5 sample    NA      NA 14.3   56     5   5
# 6 sample    28      NA 14.9   66     5   6

dat <- bind_rows(dat_sample, dat_summary)

ft <- flextable(dat, col_keys = colnames(airquality)) %>% 
  colformat_double(digits = 2, j = colnames(airquality)) %>% 
  theme_booktabs() %>% 
  bold(i = nrow_part(., part = "body")) %>% 
  footnote(i = nrow_part(., part = "body"), j = 1:4, 
           value = as_paragraph("Calculated mean"), 
           ref_symbols = c("(1)")) %>% 
  set_caption("a *weird* table summary")

ft

Table 11.1: a *weird* table summary
Ozone	Solar.R	Wind	Temp	Month	Day
41.00	190.00	7.40	67.00	5	1
36.00	118.00	8.00	72.00	5	2
12.00	149.00	12.60	74.00	5	3
18.00	313.00	11.50	62.00	5	4
<na>	<na>	14.30	56.00	5	5
28.00	<na>	14.90	66.00	5	6
42.13(1)	185.93(1)	9.96(1)	77.88(1)	<na>	<na>
(1)Calculated mean

11.2 Function ‘before’

hline(ft, i = ~ before(label, "summary"), 
        border = fp_border_default(width = 2))

Table 11.2: a *weird* table summary
Ozone	Solar.R	Wind	Temp	Month	Day
41.00	190.00	7.40	67.00	5	1
36.00	118.00	8.00	72.00	5	2
12.00	149.00	12.60	74.00	5	3
18.00	313.00	11.50	62.00	5	4
<na>	<na>	14.30	56.00	5	5
28.00	<na>	14.90	66.00	5	6
42.13(1)	185.93(1)	9.96(1)	77.88(1)	<na>	<na>
(1)Calculated mean

11.3 Looping over columns with compose

Function compose() (or mk_par()) can be used to define content of several columns in one call. Argument value can be a single as_paragraph() call if use_dot=TRUE. In that case, the call need to refer to the column value as . and not with its name. That way, multiple columns can have their content updated with one single call.

If use_dot=TRUE, value is evaluated within a data.frame augmented of a column named . containing the jth column.

The following illustration is made of two steps. The first one is only data transformation, we process the data for flextable by aggregating, plotting and pivoting:

dat <- nest(diamonds, data = -all_of(c("cut", "color"))) %>% 
  mutate(
    gg = 
      lapply(
        X = data, 
        FUN = function(subdat) {
          ggplot(subdat, aes(x = x)) + 
            geom_density(color = "white") + theme_minimal() +
            scale_x_continuous(limits = c(0, 11)) +
            scale_y_continuous(limits = c(0, 1)) +
            labs(x = "", y = "") + theme_void()
        }
      )
  ) %>% 
  select(-data) %>% 
  pivot_wider(
    id_cols = cut, 
    names_from = color, 
    values_from = gg)

dat

# # A tibble: 5 × 8
#   cut       E      I      J      H      F      G      D     
#   <ord>     <list> <list> <list> <list> <list> <list> <list>
# 1 Ideal     <gg>   <gg>   <gg>   <gg>   <gg>   <gg>   <gg>  
# 2 Premium   <gg>   <gg>   <gg>   <gg>   <gg>   <gg>   <gg>  
# 3 Good      <gg>   <gg>   <gg>   <gg>   <gg>   <gg>   <gg>  
# 4 Very Good <gg>   <gg>   <gg>   <gg>   <gg>   <gg>   <gg>  
# 5 Fair      <gg>   <gg>   <gg>   <gg>   <gg>   <gg>   <gg>

The second step is now straightforward, it format all columns but ‘cut’ as ggplot of size “1 cm * 1 cm”. Note the usage of use_dot = TRUE and . in the call to gg_chunk().

dat %>% 
  flextable() %>% 
  mk_par(
    value = as_paragraph(
      gg_chunk(., width = 1, height = 1, unit = "cm")), 
    j = ~ . - cut,
    use_dot = TRUE) %>% 
  theme_tron() %>% 
  align(align = "center", part = "all") %>% 
  autofit()

cut	E	I	J	H	F	G	D
Ideal
Premium
Good
Very Good
Fair