Code for finding the middle letter of a name in "babynames" dataset is giving faulty result when the name has 6 letters

I am trying to solve an exercise question from R for data science (2E) of chapter 16 (16.5.4 question no. 1) which requires to extract the middle letter of every name of the dataset. So I wrote the code below to find the middle letter if the name has odd number of letters or the middle two letters if the name has even number of letter.

 library(tidyverse)
 library(babynames)

   babynames |>
   mutate(
    length = str_length(name),
    middle = if_else((length / 2) %% 2 != 0, 
             str_sub(name, ceiling(length / 2), ceiling(length / 2)),
             str_sub(name, length / 2, (length / 2)+1)
    )
   )

Now the code gives me my expected result except when the name has 6 letters. Instead of extracting the middle two letters it shows only the first of the two letters

    # A tibble: 1,924,665 × 7
    year sex   name          n   prop length middle
   <dbl> <chr> <chr>     <int>  <dbl>  <int> <chr> 
 1  1880 F     Mary       7065 0.0724      4 ar    
 2  1880 F     Anna       2604 0.0267      4 nn    
 3  1880 F     Emma       2003 0.0205      4 mm    
 4  1880 F     Elizabeth  1939 0.0199      9 a     
 5  1880 F     Minnie     1746 0.0179      6 n     
 6  1880 F     Margaret   1578 0.0162      8 ga    
 7  1880 F     Ida        1472 0.0151      3 d     
 8  1880 F     Alice      1414 0.0145      5 i     
 9  1880 F     Bertha     1320 0.0135      6 r     
10  1880 F     Sarah      1288 0.0132      5 r     
# … with 1,924,655 more rows
# ℹ Use `print(n = ...)` to see more rows

I don't understand why the code is making an exception for the names with 6 letters. What can be the reason behind this?

Your check for odd/even values is incorrect. Look at

data.frame(length=1:10) |>
  mutate(compare=(length / 2) %% 2 != 0)
#    length compare
# 1       1    TRUE
# 2       2    TRUE
# 3       3    TRUE
# 4       4   FALSE
# 5       5    TRUE
# 6       6    TRUE
# 7       7    TRUE
# 8       8   FALSE
# 9       9    TRUE
# 10     10    TRUE

Notice that's not triggering correctly for even/odd values. The extra /2 in there is checking that numbers are actually divisible by 4. You should be using

babynames |>
  mutate(
    length = str_length(name),
    middle = if_else(length %% 2 != 0, 
                     str_sub(name, ceiling(length / 2), ceiling(length / 2)),
                     str_sub(name, length / 2, (length / 2)+1)
    )
  )
#     year sex   name          n   prop length middle
#    <dbl> <chr> <chr>     <int>  <dbl>  <int> <chr> 
#  1  1880 F     Mary       7065 0.0724      4 ar    
#  2  1880 F     Anna       2604 0.0267      4 nn    
#  3  1880 F     Emma       2003 0.0205      4 mm    
#  4  1880 F     Elizabeth  1939 0.0199      9 a     
#  5  1880 F     Minnie     1746 0.0179      6 nn    
#  6  1880 F     Margaret   1578 0.0162      8 ga   
# ...