Reorder rows using custom order

An alternative could be the following, note that arrange can take a list of new indices.

library(dplyr)

df <- data.frame(category = LETTERS[1:3], b = 1:3, stringsAsFactors = F)

# create a vector with letters in the desired order
x <- c("C", "A", "B")

df %>%
arrange(sapply(category, function(y) which(y == x))

First, create a vector with the letters in the desired order. Then match* the vector with the variable to be sorted. match returns indices of (first) matches, which can be plugged into slice:

library(dplyr)

# create a vector with letters in the desired order
x <- c("C", "A", "B")

DT %>%
  slice(match(x, category))
#   category b
# 1        C 3
# 2        A 1
# 3        B 2

Another way would be to convert "category" to a factor, set levels to the desired order, and use arrange:

DT %>%
  mutate(category =  factor(category, levels = x)) %>%
  arrange(category)    
#   category b
# 1        C 3
# 2        A 1
# 3        B 2

*The match method is inspired by this answer.