For loop breaks condformat [closed]

Question

A strange behavior:

data(iris)
library(condformat)
#library(knitr)
df <- iris[c(1:5,70:75, 120:125),]

cd <- condformat(df)

cd <- rule_text_bold(cd, 2, df[,2] == apply(df[,2:3], 1, FUN=min))
cd <- rule_text_bold(cd, 3, df[,3] == apply(df[,2:3], 1, FUN=min))

cd

Produces the correct output with the minimum of rows 2 and 3 in bold.

But when I switch to a for loop:

data(iris)
library(condformat)
#library(knitr)
df <- iris[c(1:5,70:75, 120:125),]

cd <- condformat(df)

for (j in 2:3){
  cd <- rule_text_bold(cd, j, df[,j] == apply(df[,2:3], 1, FUN=min))
}
cd

It breaks. Only one of the columns is highlighted. Why? I need to do this for large tables and need the loop!

Answer 1

Maybe this question would belong better in stackoverflow.

Anyway, the way condformat works is by capturing the expressions and evaluating them when the rendering is needed, similarly to what ggplot2 does. This is done using a technique called Non Standard Evaluation (NSE), leveraging on the rlang package.

You can read more about NSE here: https://dplyr.tidyverse.org/articles/programming.html

rlang provides an unquote operator called "bang-bang" !!. This operator allows the user to unquote parts of the expression, to replace variables with their values.

See for instance how the !! operator is used in your example below to replace the value of j:

data(iris)
library(condformat)
#library(knitr)
df <- iris[c(1:5,70:75, 120:125),]

cd <- condformat(df)

for (j in 2:3){
  cd <- rule_text_bold(cd, !!j, df[,!!j] == apply(df[,2:3], 1, FUN=min))
}
cd

On condformat 0.8.0, you may see a warning once per session due to recent updates of the rlang package. The warning won't appear in the next version (currently at GitHub).