Generating Random Strings - r

Question

I want to generate random strings in the following way: ABCDE1234E, i.e each string contains 5 Characters, 4 Numerics, then 1 Char.
I figured out a way to create this using the following code.
library(random)
string_5 <- as.vector(randomStrings(n=5000, len=5, digits=FALSE, upperalpha=TRUE,
loweralpha=FALSE, unique=TRUE, check=TRUE))
number_4 <- as.vector(randomNumbers(n=5000, min=1111, max=9999, col=5, base=10, check=TRUE))
string_1 <- as.vector(randomStrings(n=5000, len=1, digits=FALSE, upperalpha=TRUE,
loweralpha=FALSE, unique=FALSE, check=TRUE))
PAN.Number <- paste(string_5,number_4,string_1,sep = "")
But these functions are taking a long time and the random library needs a network connection.
> system.time(string_5 <- as.vector(randomStrings(n=5000, len=5, digits=FALSE, upperalpha=TRUE,
+ loweralpha=FALSE, unique=TRUE, check=TRUE)))
user system elapsed
0.07 0.00 3.18
Is there any method that I could try to reduce the execution time?
I also tried using sample() but I couldn't figure it out.

Answer 1

Using "stringi" as suggested by #akrun will be faster, but the following is also very fast and does not require any additional packages:
myFun <- function(n = 5000) {
a <- do.call(paste0, replicate(5, sample(LETTERS, n, TRUE), FALSE))
paste0(a, sprintf("%04d", sample(9999, n, TRUE)), sample(LETTERS, n, TRUE))
}
Example output:
myFun(10)
## [1] "BZHOF3737P" "EPOWI0674X" "YYWEB2825M" "HQIXJ5187K" "IYIMB2578R"
## [6] "YSGBG6609I" "OBLBL6409Q" "PUMAL5632D" "ABRAT4481L" "FNVEN7870Q"

Answer 2

We can use stri_rand_strings from stringi
library(stringi)
sprintf("%s%s%s", stri_rand_strings(5, 5, '[A-Z]'),
stri_rand_strings(5, 4, '[0-9]'), stri_rand_strings(5, 1, '[A-Z]'))
Or more compactly
do.call(paste0, Map(stri_rand_strings, n=5, length=c(5, 4, 1),
pattern = c('[A-Z]', '[0-9]', '[A-Z]')))
Benchmarks
system.time({
do.call(paste0, Map(stri_rand_strings, n=5000, length=c(5, 4, 1),
pattern = c('[A-Z]', '[0-9]', '[A-Z]')))
})
# user system elapsed
# 0 0 0
Was able to reproduce the timings even for one part of the expected output using OP's method
system.time(string_5 <- as.vector(randomStrings(n=5000, len=5, digits=FALSE, upperalpha=TRUE,
loweralpha=FALSE, unique=TRUE, check=TRUE)))
# user system elapsed
# 0.86 0.24 5.52

Answer 3

You can directly perform what you want:
Sample random 5 capital letters
Sample 4 digits
Sample 1 random capital letter
digits = 0:9
createRandString<- function() {
v = c(sample(LETTERS, 5, replace = TRUE),
sample(digits, 4, replace = TRUE),
sample(LETTERS, 1, replace = TRUE))
return(paste0(v,collapse = ""))
}
This will be more easily controlled, and won't take as long.

Answer 4

Your performance problem comes from using the random package in the first place: it's understandable that you could find the random::randomStrings() function in an internet search and think it's a good way to generate random strings for use in a program, but the random package is not intended for general-purpose programming. It works by querying the RANDOM.ORG server, which is intrinsically slower than R's built-in pseudo-random number generators.
From one of the vignettes from the random package:
There are a number of situations in which it is desirable to use non-deterministically determined
random numbers. Examples include
- to seed distributed computing on different nodes with truly indepedent seeds;
- to obtain portable initializations for RNGs that do not depend on particular operating system
or hardware features;
- to validate simulation results using non-deterministic random numbers;
- to provide indeterministic seeds used for lottery drawings or games ...
Note that most of these examples are about seeding or initializing (these are synonyms) R's built-in pseudo-random number generators, rather than replacing them ...

Answer 5

In case anyone came here looking for a way to generate random file names, here's what I used. I like it for its elegance
library(dplyr)
runif(1, 1000000000000, 9999999999999) %>% round %>% as.character %>% paste0("/tmp/", ., ".png")
Note: you can easily change how many random strings it generates by changing the 1 in runif() to the number you want

Answer 6

You can use the ASCII table to get a fine control of your final string.
randString <- function(characters=0, numbers=0, symbols=0, lowerCase=0, upperCase=0) {
ASCII <- NULL
if(symbols>0) ASCII <- c(ASCII, sample(c(33:47, 58:34, 91:96, 123:126), symbols))
if(numbers>0) ASCII <- c(ASCII, sample(48:57, numbers))
if(upperCase>0) ASCII <- c(ASCII, sample(65:90, upperCase))
if(lowerCase>0) ASCII <- c(ASCII, sample(97:122, lowerCase))
if(characters>0) ASCII <- c(ASCII, sample(c(65:90, 97:122), characters))
return( rawToChar(as.raw(sample(ASCII, length(ASCII)))) )
}
Example:
randString(characters=5, numbers=4)
# [1] "9fKW75o1N"

Answer 7

We can now do this with "rowwise" and "mutate" from dplyr, with library(stringi) for the stri_rand_strings function:
df %>%
rowwise() %>%
mutate(unique_id = paste0(stri_rand_strings(1, 5, "[A-Z]"), stri_rand_strings(1, 4, "[0-9]"), stri_rand_strings(1, 1, "[A-Z]")))
This avoids the need to create a function.