macroexpand doesn't require any kind of quoting

user=> (macroexpand (list  (symbol "+") 1 2))
(+ 1 2)
user=>

I am using (bootstrapped) clojurescript.

I see that clojure proper's macro expander would be okay with this.

SIgh, thanks.

You can write your code normally and wrap it in a delay Why do you want to do it lazily?

qp=> (defn get-char [number] (let [s (str number)] (fn [n] (nth s n))))
#'nocommit.qp/get-char
qp=> (let [f (get-char 9876)]
       (map f [0 1 2 3]))
(\9 \8 \7 \6)

assemble how you like but (nth (str number) index) does what you want

> Why do you want to do it lazily? To work with realy big numbers :-)

I still don’t follow. But my example works just fine

Thnx!

if you don’t want to repeatedly str the number my get-char closes over the string representationo

I don't see why the number's size matters

but otherwise index access to a string is constant (i think)

Standart function nth are not lazy.

I’m still not following. Can you explain how you would make + lazy?

the only thing i can think of is like haskell where the expression is not computed until the result is “needed”. And Clojure offers nothing like that

I'm trying to solve problems from https://projecteuler.net/ And here a lot of problems related to lazy evaluations of few last digits of some realy big number.

Guys I'm newbie, so don't be scary :-)

> the only thing i can think of is like haskell where the expression is not computed until the result is “needed”. And Clojure offers nothing like that Aren't delay and lazy sequences like that?

you can get a digit using just some math, mainly div and mod

nth digit from the end

(require '[clojure.math :as math])

(defn nth-digit [^long num ^long n]
  (mod (quot num (long (math/pow 10 n))) 10))

(nth-digit 123456 1)

For example: N equal to sum of 1 trillion of prime numbers. Find last digit of N (like ...4189 -> 9).

Thnx!

it won't work for bignums, you'll need to remove the type hints

my str version will happily work with bignums and give constant time results

isn't the str version linear time because of (str number)?

that’s done once and closed over

so depends. its amortized. but if you only get one digit then yes

right, it looks like one of the goals is to get one of the least significant digits quickly

the fun question is - is there a way to get the nth digit without calculating everything?

for example, if you need the nth digit of the sum of all primes, you can always mod the result 10^(n+1)

so you don't have to hold an insane sum of numbers

even further, is there an analytical expression for the nth digit of the sum of k primes?

i wonder what the internal representation of bignums, bigints are. It might have a string representation in there so the str call is just a constant field access and then .charAt is constant as well

probably byte arrays

i doubt there is a closed form expression for that sum

not for the sum itself, but what about a specific digit?

doubt even more but would love to hear the answer

I'm going to sleep before I get nerd sniped by Reimman

haha

My question was stupid :-(

absolutely not. it was an honest question and we had a great discussion

and it led to interesting places. everyone here is better off having thought about it, wrestled with it