Largest number divisible by its digits: Difference between revisions
MaiconSoft (talk | contribs) Added Delphi example |
m →{{header|Phix}}: added syntax colouring the hard way |
||
Line 1,386: | Line 1,386: | ||
=== base 10 === |
=== base 10 === |
||
{{trans|Go}} |
{{trans|Go}} |
||
<lang Phix> |
<!--<lang Phix>(phixonline)--> |
||
<span style="color: #008080;">with</span> <span style="color: #008080;">javascript_semantics</span> |
|||
high = 9876432, |
|||
<span style="color: #004080;">integer</span> <span style="color: #000000;">magic</span> <span style="color: #0000FF;">=</span> <span style="color: #000000;">9</span><span style="color: #0000FF;">*</span><span style="color: #000000;">8</span><span style="color: #0000FF;">*</span><span style="color: #000000;">7</span><span style="color: #0000FF;">,</span> |
|||
n = high-mod(high,magic) |
|||
<span style="color: #000000;">high</span> <span style="color: #0000FF;">=</span> <span style="color: #000000;">9876432</span><span style="color: #0000FF;">,</span> |
|||
sequence seen |
|||
<span style="color: #000000;">n</span> <span style="color: #0000FF;">=</span> <span style="color: #000000;">high</span><span style="color: #0000FF;">-</span><span style="color: #7060A8;">mod</span><span style="color: #0000FF;">(</span><span style="color: #000000;">high</span><span style="color: #0000FF;">,</span><span style="color: #000000;">magic</span><span style="color: #0000FF;">)</span> |
|||
while true do |
|||
<span style="color: #004080;">sequence</span> <span style="color: #000000;">seen</span> |
|||
string s = sprintf("%d",n) |
|||
<span style="color: #008080;">while</span> <span style="color: #004600;">true</span> <span style="color: #008080;">do</span> |
|||
seen = {1,0,0,0,0,1,0,0,0,0} |
|||
<span style="color: #004080;">string</span> <span style="color: #000000;">s</span> <span style="color: #0000FF;">=</span> <span style="color: #7060A8;">sprintf</span><span style="color: #0000FF;">(</span><span style="color: #008000;">"%d"</span><span style="color: #0000FF;">,</span><span style="color: #000000;">n</span><span style="color: #0000FF;">)</span> |
|||
for j=1 to length(s) do |
|||
<span style="color: #000000;">seen</span> <span style="color: #0000FF;">=</span> <span style="color: #0000FF;">{</span><span style="color: #000000;">1</span><span style="color: #0000FF;">,</span><span style="color: #000000;">0</span><span style="color: #0000FF;">,</span><span style="color: #000000;">0</span><span style="color: #0000FF;">,</span><span style="color: #000000;">0</span><span style="color: #0000FF;">,</span><span style="color: #000000;">0</span><span style="color: #0000FF;">,</span><span style="color: #000000;">1</span><span style="color: #0000FF;">,</span><span style="color: #000000;">0</span><span style="color: #0000FF;">,</span><span style="color: #000000;">0</span><span style="color: #0000FF;">,</span><span style="color: #000000;">0</span><span style="color: #0000FF;">,</span><span style="color: #000000;">0</span><span style="color: #0000FF;">}</span> |
|||
seen[s[j]-'0'+1] += 1 |
|||
<span style="color: #008080;">for</span> <span style="color: #000000;">j</span><span style="color: #0000FF;">=</span><span style="color: #000000;">1</span> <span style="color: #008080;">to</span> <span style="color: #7060A8;">length</span><span style="color: #0000FF;">(</span><span style="color: #000000;">s</span><span style="color: #0000FF;">)</span> <span style="color: #008080;">do</span> |
|||
end for |
|||
<span style="color: #000000;">seen</span><span style="color: #0000FF;">[</span><span style="color: #000000;">s</span><span style="color: #0000FF;">[</span><span style="color: #000000;">j</span><span style="color: #0000FF;">]-</span><span style="color: #008000;">'0'</span><span style="color: #0000FF;">+</span><span style="color: #000000;">1</span><span style="color: #0000FF;">]</span> <span style="color: #0000FF;">+=</span> <span style="color: #000000;">1</span> |
|||
if max(seen)=1 then exit end if |
|||
<span style="color: #008080;">end</span> <span style="color: #008080;">for</span> |
|||
n -= magic |
|||
<span style="color: #008080;">if</span> <span style="color: #7060A8;">max</span><span style="color: #0000FF;">(</span><span style="color: #000000;">seen</span><span style="color: #0000FF;">)=</span><span style="color: #000000;">1</span> <span style="color: #008080;">then</span> <span style="color: #008080;">exit</span> <span style="color: #008080;">end</span> <span style="color: #008080;">if</span> |
|||
end while |
|||
<span style="color: #000000;">n</span> <span style="color: #0000FF;">-=</span> <span style="color: #000000;">magic</span> |
|||
-- may as well quickly verify... |
|||
<span style="color: #008080;">end</span> <span style="color: #008080;">while</span> |
|||
seen[5+1] = 0 -- (skipping 5) |
|||
<span style="color: #000080;font-style:italic;">-- may as well quickly verify...</span> |
|||
for i=1 to 9 do -- ( and 0 ) |
|||
<span style="color: #000000;">seen</span><span style="color: #0000FF;">[</span><span style="color: #000000;">5</span><span style="color: #0000FF;">+</span><span style="color: #000000;">1</span><span style="color: #0000FF;">]</span> <span style="color: #0000FF;">=</span> <span style="color: #000000;">0</span> <span style="color: #000080;font-style:italic;">-- (skipping 5)</span> |
|||
if seen[i+1] then |
|||
<span style="color: #008080;">for</span> <span style="color: #000000;">i</span><span style="color: #0000FF;">=</span><span style="color: #000000;">1</span> <span style="color: #008080;">to</span> <span style="color: #000000;">9</span> <span style="color: #008080;">do</span> <span style="color: #000080;font-style:italic;">-- ( and 0 )</span> |
|||
if mod(n,i)!=0 then ?9/0 end if |
|||
<span style="color: #008080;">if</span> <span style="color: #000000;">seen</span><span style="color: #0000FF;">[</span><span style="color: #000000;">i</span><span style="color: #0000FF;">+</span><span style="color: #000000;">1</span><span style="color: #0000FF;">]</span> <span style="color: #008080;">then</span> |
|||
end if |
|||
<span style="color: #008080;">if</span> <span style="color: #7060A8;">mod</span><span style="color: #0000FF;">(</span><span style="color: #000000;">n</span><span style="color: #0000FF;">,</span><span style="color: #000000;">i</span><span style="color: #0000FF;">)!=</span><span style="color: #000000;">0</span> <span style="color: #008080;">then</span> <span style="color: #0000FF;">?</span><span style="color: #000000;">9</span><span style="color: #0000FF;">/</span><span style="color: #000000;">0</span> <span style="color: #008080;">end</span> <span style="color: #008080;">if</span> |
|||
end for |
|||
<span style="color: #008080;">end</span> <span style="color: #008080;">if</span> |
|||
printf(1,"%d (%d iterations)\n",{n,(high-n)/magic})</lang> |
|||
<span style="color: #008080;">end</span> <span style="color: #008080;">for</span> |
|||
<span style="color: #7060A8;">printf</span><span style="color: #0000FF;">(</span><span style="color: #000000;">1</span><span style="color: #0000FF;">,</span><span style="color: #008000;">"%d (%d iterations)\n"</span><span style="color: #0000FF;">,{</span><span style="color: #000000;">n</span><span style="color: #0000FF;">,(</span><span style="color: #000000;">high</span><span style="color: #0000FF;">-</span><span style="color: #000000;">n</span><span style="color: #0000FF;">)/</span><span style="color: #000000;">magic</span><span style="color: #0000FF;">})</span> |
|||
<!--</lang>--> |
|||
{{out}} |
{{out}} |
||
<pre> |
<pre> |
||
Line 1,415: | Line 1,418: | ||
{{trans|Haskell}} |
{{trans|Haskell}} |
||
{{libheader|Phix/mpfr}} |
{{libheader|Phix/mpfr}} |
||
<!--<lang Phix>(phixonline)--> |
|||
using gmp (15 times faster than bigatom) |
|||
<span style="color: #008080;">with</span> <span style="color: #008080;">javascript_semantics</span> |
|||
<lang Phix>atom t0 = time() |
|||
<span style="color: #004080;">atom</span> <span style="color: #000000;">t0</span> <span style="color: #0000FF;">=</span> <span style="color: #7060A8;">time</span><span style="color: #0000FF;">()</span> |
|||
integer count = 0 |
|||
<span style="color: #004080;">integer</span> <span style="color: #000000;">count</span> <span style="color: #0000FF;">=</span> <span style="color: #000000;">0</span> |
|||
string s |
|||
<span style="color: #004080;">string</span> <span style="color: #000000;">s</span> |
|||
include mpfr.e |
|||
<span style="color: #008080;">include</span> <span style="color: #004080;">mpfr</span><span style="color: #0000FF;">.</span><span style="color: #000000;">e</span> |
|||
mpz lcm15 = mpz_init(lcm(tagset(15))), |
|||
<span style="color: #004080;">mpz</span> <span style="color: #000000;">lcm15</span> <span style="color: #0000FF;">=</span> <span style="color: #7060A8;">mpz_init</span><span style="color: #0000FF;">(</span><span style="color: #7060A8;">lcm</span><span style="color: #0000FF;">(</span><span style="color: #7060A8;">tagset</span><span style="color: #0000FF;">(</span><span style="color: #000000;">15</span><span style="color: #0000FF;">))),</span> |
|||
d = mpz_init(), |
|||
<span style="color: #000000;">d</span> <span style="color: #0000FF;">=</span> <span style="color: #7060A8;">mpz_init</span><span style="color: #0000FF;">(),</span> |
|||
r = mpz_init() |
|||
<span style="color: #000000;">r</span> <span style="color: #0000FF;">=</span> <span style="color: #7060A8;">mpz_init</span><span style="color: #0000FF;">()</span> |
|||
mpz_set_str(d,"FEDCBA987654321",16) |
|||
<span style="color: #7060A8;">mpz_set_str</span><span style="color: #0000FF;">(</span><span style="color: #000000;">d</span><span style="color: #0000FF;">,</span><span style="color: #008000;">"FEDCBA987654321"</span><span style="color: #0000FF;">,</span><span style="color: #000000;">16</span><span style="color: #0000FF;">)</span> |
|||
mpz_mod(r,d,lcm15) |
|||
<span style="color: #7060A8;">mpz_mod</span><span style="color: #0000FF;">(</span><span style="color: #000000;">r</span><span style="color: #0000FF;">,</span><span style="color: #000000;">d</span><span style="color: #0000FF;">,</span><span style="color: #000000;">lcm15</span><span style="color: #0000FF;">)</span> |
|||
mpz_sub(d,d,r) -- d:= max k*lcm <= "FE..21" |
|||
<span style="color: #7060A8;">mpz_sub</span><span style="color: #0000FF;">(</span><span style="color: #000000;">d</span><span style="color: #0000FF;">,</span><span style="color: #000000;">d</span><span style="color: #0000FF;">,</span><span style="color: #000000;">r</span><span style="color: #0000FF;">)</span> <span style="color: #000080;font-style:italic;">-- d:= max k*lcm <= "FE..21"</span> |
|||
r = mpz_free(r) -- (no longer used) |
|||
<span style="color: #000000;">r</span> <span style="color: #0000FF;">=</span> <span style="color: #7060A8;">mpz_free</span><span style="color: #0000FF;">(</span><span style="color: #000000;">r</span><span style="color: #0000FF;">)</span> <span style="color: #000080;font-style:italic;">-- (no longer used)</span> |
|||
while true do |
|||
<span style="color: #008080;">while</span> <span style="color: #004600;">true</span> <span style="color: #008080;">do</span> |
|||
s = mpz_get_str(d,16) |
|||
<span style="color: #000000;">s</span> <span style="color: #0000FF;">=</span> <span style="color: #7060A8;">mpz_get_str</span><span style="color: #0000FF;">(</span><span style="color: #000000;">d</span><span style="color: #0000FF;">,</span><span style="color: #000000;">16</span><span style="color: #0000FF;">)</span> |
|||
if sort(s)="123456789abcdef" then exit end if |
|||
<span style="color: #008080;">if</span> <span style="color: #7060A8;">sort</span><span style="color: #0000FF;">(</span><span style="color: #000000;">s</span><span style="color: #0000FF;">)=</span><span style="color: #008000;">"123456789abcdef"</span> <span style="color: #008080;">then</span> <span style="color: #008080;">exit</span> <span style="color: #008080;">end</span> <span style="color: #008080;">if</span> |
|||
mpz_sub(d,d,lcm15) |
|||
<span style="color: #7060A8;">mpz_sub</span><span style="color: #0000FF;">(</span><span style="color: #000000;">d</span><span style="color: #0000FF;">,</span><span style="color: #000000;">d</span><span style="color: #0000FF;">,</span><span style="color: #000000;">lcm15</span><span style="color: #0000FF;">)</span> |
|||
count += 1 |
|||
<span style="color: #000000;">count</span> <span style="color: #0000FF;">+=</span> <span style="color: #000000;">1</span> |
|||
end while |
|||
<span style="color: #008080;">end</span> <span style="color: #008080;">while</span> |
|||
string e = elapsed(time()-t0) |
|||
<span style="color: #004080;">string</span> <span style="color: #000000;">e</span> <span style="color: #0000FF;">=</span> <span style="color: #7060A8;">elapsed</span><span style="color: #0000FF;">(</span><span style="color: #7060A8;">time</span><span style="color: #0000FF;">()-</span><span style="color: #000000;">t0</span><span style="color: #0000FF;">)</span> |
|||
printf(1,"%s (%d iterations, %s)\n",{s,count,e})</lang> |
|||
<span style="color: #7060A8;">printf</span><span style="color: #0000FF;">(</span><span style="color: #000000;">1</span><span style="color: #0000FF;">,</span><span style="color: #008000;">"%s (%d iterations, %s)\n"</span><span style="color: #0000FF;">,{</span><span style="color: #000000;">s</span><span style="color: #0000FF;">,</span><span style="color: #000000;">count</span><span style="color: #0000FF;">,</span><span style="color: #000000;">e</span><span style="color: #0000FF;">})</span> |
|||
<!--</lang>--> |
|||
{{out}} |
{{out}} |
||
<pre> |
<pre> |
||
fedcb59726a1348 (954460 iterations, 4s) |
fedcb59726a1348 (954460 iterations, 12.4s) |
||
</pre> |
</pre> |
||
Revision as of 07:44, 11 June 2021
![Task](http://static.miraheze.org/rosettacodewiki/thumb/b/ba/Rcode-button-task-crushed.png/64px-Rcode-button-task-crushed.png)
You are encouraged to solve this task according to the task description, using any language you may know.
- Task
Find the largest base 10 integer whose digits are all different, and is evenly divisible by each of its individual digits.
These numbers are also known as Lynch-Bell numbers, numbers n such that the
(base ten) digits are all different (and do not include zero) and n is divisible by each of its individual digits.
- Example
135 is evenly divisible by 1, 3, and 5.
Note that the digit zero (0) can not be in the number as integer division by zero is undefined.
The digits must all be unique so a base ten number will have at most 9 digits.
Feel free to use analytics and clever algorithms to reduce the search space your example needs to visit, but it must do an actual search. (Don't just feed it the answer and verify it is correct.)
- Stretch goal
Do the same thing for hexadecimal.
- Related tasks
- Also see
-
- The OEIS sequence: A115569: Lynch-Bell numbers.
11l
Base 10
<lang 11l>F check_dec(num)
Set[Int] st L(c) String(num) V d = Int(c) I d == 0 | num % d != 0 | d C st R 0B st.add(d) R 1B
L(i) (98764321 .< 0).step(-1)
I check_dec(i) print(i) L.break</lang>
- Output:
9867312
360 Assembly
For maximum compatibility, this program uses only the basic instruction set. The program uses also HLASM structured macros (DO,ENDDO,IF,ELSE,ENDIF) for readability and one ASSIST/360 macros XPRNT) to keep the code as short as possible.
Base 10
<lang 360asm>* Largest number divisible by its digits - base10 - 05/05/2020 LANUDIDO CSECT
USING LANUDIDO,R13 base register B 72(R15) skip savearea DC 17F'0' savearea SAVE (14,12) save previous context ST R13,4(R15) link backward ST R15,8(R13) link forward LR R13,R15 set addressability L R6,KM i=km DO WHILE=(C,R6,GE,=F'1') do i=km to 1 by -ks CVD R6,DBLI binary to packed decimal MVC CI,MAK load mask for edit EDMK CI,DBLI+2 ci=cstr(i) S R1,=A(CI) number of blanks LA R2,L'CI length(ci) SR R2,R1 length of the number ST R2,LCI lci=length(ci) MVC CIL,=CL12' ' cil=' ' LA R4,CI @ci AR R4,R1 @ci[k] LA R3,CIL @cil LR R5,R2 length(ci) BCTR R5,0 ~ EX R5,MVCR3R4 cil=ltrim(ci) LA R8,CIL @ci LA R7,1 j=1 DO WHILE=(C,R7,LE,LCI) do j=1 to length(ci) CLI 0(R8),C'5' if ci[j]='5' BE ITERI then cycle i ! 5 impossible CLI 0(R8),C'0' if ci[j]='0' BE ITERI then cycle i ! 0 impossible LA R8,1(R8) @ci++ LA R7,1(R7) j++ ENDDO , enddo j L R2,LCI length(ci) BCTR R2,0 length(ci)-1 LA R7,1 j=1 DO WHILE=(CR,R7,LE,R2) do j=1 to length(ci)-1 LA R3,CIL-1 @cil AR R3,R7 @cil[j] IC R1,0(R3) cil[j] STC R1,CK ck=substr(cil,j,1) SR R0,R0 index=0 LR R8,R7 j LA R8,1(R8) k=j+1 LA R9,CIL @ci AR R9,R8 +i BCTR R9,0 -1 DO WHILE=(C,R8,LE,LCI) do k=j+1 to length(ci) IF CLC,0(0,R9),EQ,CK THEN if substr(ci,k,1)=ck then LR R0,R8 index=k B EXITK exit do k ENDIF , endif LA R9,1(R9) @ci++ LA R8,1(R8) k++ ENDDO , enddo k
EXITK LTR R0,R0 if index(ci,ck,j+1)<>0
BNZ ITERI then cycle i ! no dup digit LA R7,1(R7) j++ ENDDO , enddo j LA R7,1 j=1 DO WHILE=(C,R7,LE,LCI) do j=1 to length(ci) LA R3,CIL-1 @cil AR R3,R7 @cil[j] IC R1,0(R3) cil[j] SLL R1,28 ~ SRL R1,28 kk=int(substr(ci,j,1)) XR R4,R4 ~ LR R5,R6 i DR R4,R1 r5=i/kk r4=mod(i,kk) LTR R4,R4 if mod(i,kk)<>0 BNZ ITERI then cycle i ! div by all digit LA R7,1(R7) j++ ENDDO , enddo j B EXITI exit do i ! found
ITERI S R6,KS i-=ks
ENDDO , enddo i
EXITI XPRNT CIL,L'CIL print cil
L R13,4(0,R13) restore previous savearea pointer RETURN (14,12),RC=0 restore registers from calling save
MVCR3R4 MVC 0(0,R3),0(R4) move: %R3 <- %R4 KS DC F'504' ks=504=7*8*9 magic number KM DC F'9876384' km=9876384 mod(km,ks)=0 DBLI DS PL8 double word 15num MAK DC X'402020202020202020202020' mask CL12 11num CI DS CL12 12num - i right justify CIL DS CL12 12num - i left justify LCI DS F length(i) CK DS C ck
REGEQU END LANUDIDO </lang>
- Output:
9867312
AWK
Base 10
Ruling out 9- and 8-digit numbers (see first paragraph in the Raku example), we are looking for 7-digit numbers. In order to be a solution such a number has to be divisible by 12 = 2*2*3, because its digits must contain at least 2 of the numbers 2, 4, 6, 8 (leading to a factor of 2*2) and its digits must contain at least one of the numbers 3, 6, 9 (leading to a factor of 3).
The program does a brute force search, starting with the largest possible 7-digit number and iterates over all smaller numbers divisible by 12. It checks for each iteration, if the number in question consists of different digits and is divisible by those digits. <lang AWK># Usage: gawk -f LARGEST_NUMBER_DIVISIBLE_BY_ITS_DIGITS_10.AWK BEGIN {
base = 10 comdiv = 12 startn = 9876543 stopn = 1000000 solve(startn, stopn)
} function solve(startn, stopn, n, d) {
for (n = startn - startn % comdiv; n > stopn; n -= comdiv) { if (hasuniqedigits(n)) { # Check divisibility of n by all its digits for (d = 2; d < base; d++) { if ((dcount[d]) && (n % d)) { break } } if (d == base) { printf("%d\n", n) return } } }
} function hasuniqedigits(n, d) {
# Returns 1, if n consists of unique digits in range 1..(base-1) # The array dcount stores the count (up to 1) of those digits for (d = 1; d < base; d++) dcount[d] = 0 while (n) { d = n % base if ((d == 0) || (++dcount[d] > 1)) return 0 n = int(n / base) } return 1
}</lang>
- Output:
9867312
Base 16
In the hexadecimal case we cannot rule out 15-digit numbers, thus all digits from 1 to f (hex) are present. The number has to be divisible by all its digits, therefore it has to be divisible by the least common multiple of the numbers 1, 2, 3, ..., 15 (360360).
AWK does not support arbitrary long integers, so we have to use an array of digits for its representation. It makes use of functions hexmod (modulus) and hexsub (subtraction), which act on an array.
The program does a brute force search, starting with the largest possible 15-digit number and iterates over all smaller numbers divisible by 360360. It checks for each iteration, if the number in question consists of different digits (by construction it is then also divisible by its digits).
<lang AWK># Usage: GAWK -f LARGEST_NUMBER_DIVISIBLE_BY_ITS_DIGITS_16.AWK BEGIN {
base = 16 size = 15 # startn = FEDCB A9876 54321 (hex) for (i = 1; i <= size; i++) { startn[i] = i } comdiv = 360360 # lcm(1..15) solve(startn)
} function solve(n, r, i) {
r = hexmod(n, comdiv) hexsub(n, r) while (n[size] > 0) { if (hasuniqedigits(n)) { for (i = size; i > 0; i--) printf("%0x", n[i]) printf("\n") return } hexsub(n, comdiv) }
} function hasuniqedigits(n, d, i) {
# Return 1, if n is an array of unique digits in range 1..(base-1) # The array dcount stores the count (up to 1) of those digits for (d = 1; d < base; d++) dcount[d] = 0 for (i = 1; i <= size; i++) { d = n[i] if ((d == 0) || (++dcount[d] > 1)) return 0 } return 1
} function hexmod(n, k, i, r) {
# Return n mod k, where n is an array and k is a number for (i = size; i > 0; i--) { r = (r * base + n[i]) % k } return r
} function hexsub(n, m) {
# Calculate n = n - m, where n is an array and m is a number for (i = 1; m && (i <= size); i++) { n[i] -= m % base m = int(m / base) if (n[i] < 0) { n[i] += base m++ } }
}</lang>
- Output:
fedcb59726a1348
C
Base 10
The number can't contain 0 and 5, 0 is obvious, 5 because the number must end in 5 for it to be a multiple of that number and if that happens, all the even digits are ruled out which severely reduces the number's length since the other condition is that all digits must be unique. However, this means the number must be even and thus end only in 2,4,6,8. This speeds up the search by a factor of 2. The same approach when applied to hexadecimals takes a very long, long time. <lang C>
- include<stdio.h>
int main() { int num = 9876432,diff[] = {4,2,2,2},i,j,k=0; char str[10];
start:snprintf(str,10,"%d",num);
for(i=0;str[i+1]!=00;i++){ if(str[i]=='0'||str[i]=='5'||num%(str[i]-'0')!=0){ num -= diff[k]; k = (k+1)%4; goto start; } for(j=i+1;str[j]!=00;j++) if(str[i]==str[j]){ num -= diff[k]; k = (k+1)%4; goto start; } }
printf("Number found : %d",num); return 0; } </lang> Output:
Number found : 9867312
Base 16
<lang C>#include<stdio.h>
- include<string.h>
- define TRUE 1
- define FALSE 0
typedef char bool;
typedef unsigned long long uint64;
bool div_by_all(uint64 num, char digits[], int len) {
int i, d; for (i = 0; i < len; ++i) { d = digits[i]; d = (d <= '9') ? d - '0' : d - 'W'; if (num % d != 0) return FALSE; } return TRUE;
}
int main() {
uint64 i, magic = 15 * 14 * 13 * 12 * 11; uint64 high = 0xfedcba987654321 / magic * magic; int j, len; char c, *p, s[17], sd[16], found[16];
for (i = high; i >= magic; i -= magic) { if (i % 16 == 0) continue; // can't end in '0' snprintf(s, 17, "%llx", i); // always generates lower case a-f if (strchr(s, '0') - s >= 0) continue; // can't contain '0' for (j = 0; j < 16; ++j) found[j] = FALSE; len = 0; for (p = s; *p; ++p) { if (*p <= '9') { c = *p - '0'; } else { c = *p - 87; } if (!found[c]) { found[c] = TRUE; sd[len++] = *p; } } if (len != p - s) { continue; // digits must be unique } if (div_by_all(i, sd, len)) { printf("Largest hex number is %llx\n", i); break; } } return 0;
}</lang>
- Output:
Largest hex number is fedcb59726a1348
C#
Base 10
<lang csharp>using System; using System.Collections.Generic; using System.Linq;
namespace LargestNumber {
class Program { static bool ChkDec(int num) { HashSet<int> set = new HashSet<int>();
return num.ToString() .Select(c => c - '0') .All(d => (d != 0) && (num % d == 0) && set.Add(d)); }
static void Main() { int result = Enumerable.Range(0, 98764321) .Reverse() .Where(ChkDec) .First(); Console.WriteLine(result); } }
}</lang>
- Output:
9867312
C++
Base 10
<lang cpp>#include <iostream>
- include <sstream>
- include <set>
bool checkDec(int num) {
std::set<int> set;
std::stringstream ss; ss << num; auto str = ss.str();
for (int i = 0; i < str.size(); ++i) { char c = str[i]; int d = c - '0'; if (d == 0) return false; if (num % d != 0) return false; if (set.find(d) != set.end()) { return false; } set.insert(d); }
return true;
}
int main() {
for (int i = 98764321; i > 0; i--) { if (checkDec(i)) { std::cout << i << "\n"; break; } }
return 0;
}</lang>
- Output:
9867312
Clojure
Base Agnostic
This is a generic solution that works for any number base. Just change the line (def the_base 16). The performance may be questionable for large bases which do not have a Lynch-Bell number using all digits.
Rather than searching through all numbers with unique digits, a space whose size verges on N factorial, this algorithm instead works with the non-empty sets of non-zero digits, a space of size 2^(N-1) - 1. For a given subset, it finds the least common multiple for that subset and examines each multiple of the LCM which is between the largest and smallest positive numbers that can be constructed using each digit from that subset exactly once. <lang clojure> (require '[clojure.string :as str]) ;' (def the_base 16)
- A sequence named digits containing the non-zero digits for the current number base.
(def digits (rest (range the_base)))
- A container for the digits which are prime.
(def primes [])
- Populate the primes sequence with the primes less than the current base.
(for [n digits] (if (= 1 (count (filter (fn[m] (and (< m n) (= 0 (mod n m)))) digits)) ) (def primes (conj primes n))))
- Determines the highest power of a given prime p that divides a given integer n.
(defn duplicity [n p partial] (if (= 0 (mod n p)) (duplicity (/ n p) p (conj partial p)) partial))
- Constructs the prime factorization of a given integer.
(defn factorize [n] (let [a (flatten (for [p (filter #(< % n) primes)]
(remove #(= 1 %) (duplicity n p [1]))))] (if (= 0 (count a)) (lazy-seq [n]) a) ))
- Determines the number of times a given number appears in a given sequence of numbers.
(defn multiplicity [s n] (count (filter #(= n %) s)))
- Combines two sequence two create their "union" in the sense that in the resulting sequence
- each element from each sequence is uniquely represented and no smaller sequence would suffice.
- For example if one sequence contains two A's and other contains three A's, then the result will contain three A's.
- This is used to generate representations of prime factorizations and to construct least common multiples from them.
(defn combine [x y] (concat x (flatten (for [w (dedupe y)] (repeat (- (multiplicity y w) (multiplicity x w)) w) ))))
- deterimes the lcm least common multiple for a set of digits.
(defn lcm [s] (reduce * (reduce combine (map factorize s))))
- Retuns x^n.
(defn exp [x n] (reduce * (repeat n x)))
- Generates all non-empty subsequences for a sequence.
(defn non_empty_subsets [s] (for [x (reverse (rest (range (exp 2 (count s)))))]
(remove nil? (for [i (range (count s))] (if (bit-test x i) (nth s i))))))
- Generates from a given sequence of digits in the current base the number that is s[0]s[1]s[2]...s[n].
- More generally, produces s[0]*the_base^n + s[1]*the_base^(n-1) + ... + s[n-1]*the_base^1 + s[n]*the_base^0
- for an arbitrary sequence of numbers.
(defn power_up [s] (reduce + (loop [idx (- (count s) 1) s_next s]
(if (zero? idx) s_next (recur (dec idx) (map-indexed #(if (< %1 idx) (* %2 the_base) %2) s_next))))))
- Here is an alternative version of power_up that could be more efficient as it does not repeatedly recalculate powers of the base.
- Instead it calculates the dot product of s with a pre-populated sequence of powers of the base.
- Calculates the dot product of two vectors/sequences
- (defn dot [xs ys] (reduce + (map * xs ys)))
- (def places (map #(exp the_base %) (range the_base)))
- (defn power_up [s] (dot s (reverse (take (count s) places))))
- Returns the largest integer which contains each item from a given sequence exactly once as a digit.
(defn max_for_digits [s] (power_up (sort #(> %1 %2) s)))
- Returns the smallest non-negative integer which contains each item from a given sequence exactly once as a digit.
(defn min_for_digits [s] (power_up (sort #(< %1 %2) s)))
- calculate the logarithm of the input in the current base.
(defn log_base [x] (/ (Math/log x) (Math/log the_base)))
- Removes the zeros from a sequence
(defn remove_zeros [s] (remove #(= % 0) s))
- Returns the largest integer that is a multiple of a given integer and does not exceed another given integer.
(defn first_multiple_not_after [n ub] (loop [m ub] (if (= 0 (mod m n)) m (recur (dec m)))))
- creates a representation in the current base of a positive integer as a sequence listing the digits for the number in the base.
(defn representation [n] (let [full_power (int (log_base n))]
(loop [power full_power place (exp the_base full_power) rep [] rem n ] (if (= power -1) rep (recur (dec power) (/ place the_base) (conj rep (int (/ rem place))) (- rem (* place (int (/ rem place)))))))))
- determines if a given number is exactly comprised of a given set of digits.
(defn digit_qualifies [m s] (let [rep_m (representation m)] (= (sort s) (sort rep_m))))
- Returns a sequence containing the largest Lynch-Bell number for the current base and a given sequence of digits
- or an empty sequence if there is none.
(defn find_s_largest_lb [s] (let [lb (min_for_digits s)] (let [m (lcm s)]
(loop [v (first_multiple_not_after m (max_for_digits s))] (if (< v lb) [] (if (digit_qualifies v s) (representation v) (recur (- v m))))))))
- Finds the largest Lynch-Bell number for the current base by looking for the largest for all subsets of a given size
- and picking the largest from those working from the largest size (most digits) to the smallest.
(defn find_largest_lb [] (let [subsets (non_empty_subsets (reverse digits))]
(loop [s_size (- the_base 1)] (let [hits (remove #(= (count %) 0) (map find_s_largest_lb (filter #(= (count %) s_size) subsets)))] (if (pos? (count hits)) (first (sort #(first (remove_zeros (map - %2 %1))) hits)) (recur (dec s_size)))))))
- Converts small integers to hexidecimal digits.
- This isn't being used but could be leveraged to make output that looks normal for base 16.
(defn hex_digit [v] (case v 15 "F" 14 "E" 13 "D" 12 "C" 11 "B" 10 "A" (str v)))
(find_largest_lb) </lang>
- Output:
[15 14 13 12 11 5 9 7 2 6 10 1 3 4 8] (base 16) [10 9 8 7 6 2 4 1 3] (base 11) [9 8 6 7 3 1 2] (base 10)
Crystal
base 10
<lang ruby>magic_number = 9*8*7 div = (9876432 // magic_number) * magic_number candidates = div.step(to: 0, by: -magic_number)
res = candidates.find do |c|
digits = c.to_s.chars.map(&.to_i) (digits & [0,5]).empty? && digits == digits.uniq
end
puts "Largest decimal number is #{res}"</lang>
- Output:
Largest decimal number is 9867312
base 16
<lang ruby>def divByAll(num, digits)
digits.all? { |digit| num % digit.to_i(16) == 0 }
end
magic = 15_i64 * 14 * 13 * 12 * 11 high = (0xfedcba987654321_i64 // magic) * magic
high.step(to: magic, by: -magic) do |i|
s = i.to_s(16) # always generates lower case a-f next if s.includes?('0') || s.chars.uniq.size != s.size # need uniq non-zero digits (puts "Largest hex number is #{i.to_s(16)}"; break) if divByAll(i, s.chars)
end</lang>
- Output:
Largest hex number is fedcb59726a1348
D
Base 10
<lang d>import std.algorithm.iteration : filter, map; import std.algorithm.searching : all; import std.conv : to; import std.range : iota; import std.stdio : writeln;
bool chkDec(int num) {
int[int] set;
return num .to!string .map!(c => c.to!int - '0') .all!(d => (d != 0) && (num % d == 0) && set[d]++ < 1);
}
auto lcm(R)(R r) {
return r.reduce!((a,b) => a * b / gcd(a,b));
}
void main() {
// base 10 iota(98764321, 0, -1) .filter!chkDec .front .writeln;
}</lang>
- Output:
9867312
Delphi
base 10
<lang Delphi> program Largest_number_divisible_by_its_digits;
{$APPTYPE CONSOLE}
uses
System.SysUtils;
type
TSetAnsiChar = set of AnsiChar;
function DivByAll(num: Integer; digits: TSetAnsiChar): Boolean; var
offset: byte;
begin
offset := ord('0'); for var d in digits do begin if (num mod (ord(d) - offset)) <> 0 then exit(false); end; Result := true;
end;
begin
var magic: Cardinal := 9 * 8 * 7; var h: Cardinal := 9876432 div magic * magic; for var i := h downto magic do begin if i mod 10 = 0 then Continue;
var s := i.tostring; if (s.indexOf('0') > -1) or (s.indexOf('5') > -1) then Continue; var digits: TSetAnsiChar := []; var isUnic := true; for var b in ansistring(s) do if not (b in digits) then Include(digits, b) else begin isUnic := false; break; end;
if not isUnic then Continue;
if DivByAll(i, digits) then begin writeln('Largest decimal number is ', i); Break; end; end; readln;
end.</lang>
base 16
<lang Delphi> program Largest_number_divisible_by_its_digits;
{$APPTYPE CONSOLE}
uses
System.SysUtils;
type
TSetAnsiChar = set of AnsiChar;
function DivByAll(num: Uint64; digits: TSetAnsiChar): Boolean; var
offset_dec, offset_hex: byte;
begin
offset_dec := ord('0'); offset_hex := ord('W');
for var digit in digits do begin var d: byte := 0; if digit <= '9' then d := ord(digit) - offset_dec else d := ord(digit) - offset_hex; if (num mod d) <> 0 then exit(false); end; Result := true;
end;
begin
var magic: Uint64 := 15 * 14 * 13 * 12 * 11; var h: Uint64 := $fedcba987654321 div magic * magic; writeln('Wait while search for'); for var i := h downto magic do begin if (i mod 16) = 0 then Continue;
var s := i.ToHexString(0).ToLower; if (s.indexOf('0') > -1) then Continue;
var digits: TSetAnsiChar := []; var isUnic := true; for var b in ansistring(s) do if not (b in digits) then Include(digits, b) else begin isUnic := false; break; end;
if not isUnic then Continue;
if DivByAll(i, digits) then begin writeln('Largest hex number is ', i.ToHexString); Break; end; end; readln;
end.</lang>
Factor
Base 10
This program works by filtering all the 8-digit permutations (of which there are only ~40,000) for all-digit-divisibility, and upon finding none, it will then generate the 7-digit combinations (of which there are 8) of the 8 possible digits, and then filter all permutations of the 8 combinations for all-digit-divisibility. Upon finding many, it will simply select the largest element which is our answer. If there hadn't been any 7-digit solutions, it would have gone down to six and then five, etc. <lang factor>USING: io kernel math math.combinatorics math.parser math.ranges sequences tools.time ; IN: rosetta-code.largest-divisible
- all-div? ( seq -- ? )
[ string>number ] [ string>digits ] bi [ mod ] with map sum 0 = ;
- n-digit-all-div ( n -- seq )
"12346789" swap <combinations> [ [ all-div? ] filter-permutations ] map concat ;
- largest-divisible ( -- str )
8 [ dup n-digit-all-div dup empty? ] [ drop 1 - ] while nip supremum ;
- largest-divisible-demo ( -- )
[ largest-divisible print ] time ;
MAIN: largest-divisible-demo</lang>
- Output:
9867312 Running time: 0.07224931499999999 seconds
Go
base 10
<lang go>package main
import (
"fmt" "strconv" "strings"
)
func divByAll(num int, digits []byte) bool {
for _, digit := range digits { if num%int(digit-'0') != 0 { return false } } return true
}
func main() {
magic := 9 * 8 * 7 high := 9876432 / magic * magic for i := high; i >= magic; i -= magic { if i%10 == 0 { continue // can't end in '0' } s := strconv.Itoa(i) if strings.ContainsAny(s, "05") { continue // can't contain '0'or '5' } var set = make(map[byte]bool) var sd []byte // distinct digits for _, b := range []byte(s) { if !set[b] { set[b] = true sd = append(sd, b) } } if len(sd) != len(s) { continue // digits must be unique } if divByAll(i, sd) { fmt.Println("Largest decimal number is", i) return } }
}</lang>
- Output:
Largest decimal number is 9867312
base 16
<lang go>package main
import (
"fmt" "strconv" "strings"
)
func divByAll(num int64, digits []byte) bool {
for _, digit := range digits { var d int64 if digit <= '9' { d = int64(digit - '0') } else { d = int64(digit - 'W') } if num%d != 0 { return false } } return true
}
func main() {
var magic int64 = 15 * 14 * 13 * 12 * 11 high := 0xfedcba987654321 / magic * magic for i := high; i >= magic; i -= magic { if i%16 == 0 { continue // can't end in '0' } s := strconv.FormatInt(i, 16) // always generates lower case a-f if strings.IndexByte(s, '0') >= 0 { continue // can't contain '0' } var set = make(map[byte]bool) var sd []byte // distinct digits for _, b := range []byte(s) { if !set[b] { set[b] = true sd = append(sd, b) } } if len(sd) != len(s) { continue // digits must be unique } if divByAll(i, sd) { fmt.Printf("Largest hex number is %x\n", i) return } }
}</lang>
- Output:
Largest hex number is fedcb59726a1348
Haskell
base 10
Using the analysis provided in the Raku (base 10) example:
<lang haskell>import Data.List (maximumBy, permutations, delete) import Data.Ord (comparing) import Data.Bool (bool)
unDigits :: [Int] -> Int unDigits = foldl ((+) . (10 *)) 0
ds :: [Int] ds = [1, 2, 3, 4, 6, 7, 8, 9] -- 0 (and thus 5) are both unworkable
lcmDigits :: Int lcmDigits = foldr1 lcm ds -- 504
sevenDigits :: Int sevenDigits = (`delete` ds) <$> [1, 4, 7] -- Dropping any one of these three
main :: IO () main =
print $ maximumBy -- Checking for divisibility by all digits (comparing (bool 0 <*> (0 ==) . (`rem` lcmDigits))) (unDigits <$> concat (permutations <$> sevenDigits))</lang>
- Output:
Test run from inside the Atom editor:
9867312 [Finished in 0.395s]
base 16
First member of a descending sequence of multiples of 360360 that uses the full set of 15 digits when expressed in hex. <lang haskell>import Data.Set (fromList) import Numeric (showHex)
lcmDigits :: Int lcmDigits = foldr1 lcm [1 .. 15] -- 360360
upperLimit :: Int upperLimit = allDigits - rem allDigits lcmDigits
where allDigits = 0xfedcba987654321
main :: IO () main =
(print . head) (filter ((15 ==) . length . fromList) $ (`showHex` []) <$> [upperLimit,upperLimit - lcmDigits .. 1])</lang>
Test run from inside the Atom editor:
"fedcb59726a1348" [Finished in 2.319s]
J
The 536 values found---all base 10 numbers that are divisible by their digits without repetition---are sorted descending, hence 9867312 is the greatest number divisible by its digits in base 10. <lang J>
Filter =: (#~`)(`:6) combinations =: <@#"1~ [: #: [: i. 2 ^ # permutations =: A.&i.~ ! f =: [: \:~ _ ". [: ; [: ({~ permutations@#)L:_1 }.@combinations test =: 0 = ([: +/ (|~ 10&#.inv))&>
test Filter f '12346789'
9867312 9812376 9782136 9781632 9723168 9718632 9678312 9617832 9617328 9283176 9278136 9237816 9231768 9182376 9176832 9176328 9163728 8973216 8912736 8796312 8731296 8617392 8367912 8312976 8219736 8176392 8163792 8123976 7921368 7916832 7916328 7892136 ... </lang> Working in base 16 using the largest possible solution also a multiple of the least common multiple, subtract the LCM until all the digits appear. <lang J>
NB. 16bfedcba987654321 loses precision and so we need to work in extended data type
[ HEX_DIGITS =: >: i. _15x
15 14 13 12 11 10 9 8 7 6 5 4 3 2 1
[ LCM =: *./ HEX_DIGITS
360360
] START =: <.&.(%&LCM)16#.HEX_DIGITS
1147797409030632360
Until =: conjunction def 'u^:(0-:v)^:_' assert 9 -: >:Until(>&8)1
test=: 0 -.@e. HEX_DIGITS e. 16&#.inv
[ SOLUTION =: -&LCM Until test START
1147797065081426760
'16b' , (16 #.inv SOLUTION) { Num_j_ , 26 }. Alpha_j_
16bfedcb59726a1348
</lang>
Java
Base 10
Using the analysis provided in the Raku (base 10) example: <lang java>public class LynchBell {
static String s = ""; public static void main(String args[]) { //Highest number with unique digits (no 0 or 5) int i = 98764321; boolean isUnique = true; boolean canBeDivided = true; while (i>0) { s = String.valueOf(i); isUnique = uniqueDigits(i); if (isUnique) { //Number has unique digits canBeDivided = testNumber(i); if(canBeDivided) { System.out.println("Number found: " + i); i=0; } } i--; } } public static boolean uniqueDigits(int i) { //returns true, if unique digits, false otherwise for (int k = 0; k<s.length();k++) { for(int l=k+1; l<s.length();l++) { if(s.charAt(l)=='0' || s.charAt(l)=='5') { //0 or 5 is a digit return false; } if(s.charAt(k) == s.charAt(l)) { //non-unique digit return false; } } } return true; } public static boolean testNumber(int i) { //Tests, if i is divisible by all its digits (0 is not a digit already) int j = 0; boolean divisible = true; // TODO: divisible by all its digits for (char ch: s.toCharArray()) { j = Character.getNumericValue(ch); divisible = ((i%j)==0); if (!divisible) { return false; } } return true; }
}</lang>
- Output:
Number found: 9867312
Julia
Base 10
<lang julia>function main()
num = 9876432 dif = [4, 2, 2, 2] local k = 1 @label start local str = dec(num) for (i, ch) in enumerate(str) if ch in ('0', '5') || num % (ch - '0') != 0 num -= dif[k] k = (k + 1) % 4 + 1 @goto start end for j in i+1:endof(str) if str[i] == str[j] num -= dif[k] k = (k + 1) % 4 + 1 @goto start end end end
return num
end
println("Number found: ", main())</lang>
- Output:
Number found: 9867312
Kotlin
Makes use of the Raku entry's analysis:
base 10
<lang scala>// version 1.1.4-3
fun Int.divByAll(digits: List<Char>) = digits.all { this % (it - '0') == 0 }
fun main(args: Array<String>) {
val magic = 9 * 8 * 7 val high = 9876432 / magic * magic for (i in high downTo magic step magic) { if (i % 10 == 0) continue // can't end in '0' val s = i.toString() if ('0' in s || '5' in s) continue // can't contain '0' or '5' val sd = s.toCharArray().distinct() if (sd.size != s.length) continue // digits must be unique if (i.divByAll(sd)) { println("Largest decimal number is $i") return } }
}</lang>
- Output:
Largest decimal number is 9867312
base 16
<lang scala>// version 1.1.4-3
fun Long.divByAll(digits: List<Char>) =
digits.all { this % (if (it <= '9') it - '0' else it - 'W') == 0L }
fun main(args: Array<String>) {
val magic = 15L * 14 * 13 * 12 * 11 val high = 0xfedcba987654321L / magic * magic for (i in high downTo magic step magic) { if (i % 16 == 0L) continue // can't end in '0' val s = i.toString(16) // always generates lower case a-f if ('0' in s) continue // can't contain '0' val sd = s.toCharArray().distinct() if (sd.size != s.length) continue // digits must be unique if (i.divByAll(sd)) { println("Largest hex number is ${i.toString(16)}") return } }
}</lang>
- Output:
Largest hex number is fedcb59726a1348
Lua
<lang lua>function isDivisible(n)
local t = n local a = {0, 0, 0, 0, 0, 0, 0, 0, 0, 0}
while t ~= 0 do local r = t % 10 if r == 0 then return false end if n % r ~= 0 then return false end if a[r + 1] > 0 then return false end a[r + 1] = 1 t = math.floor(t / 10) end
return true
end
for i=9999999999,0,-1 do
if isDivisible(i) then print(i) break end
end </lang>
- Output:
9867312
Mathematica
base 10
<lang Mathematica> Max@Select[FromDigits/@Rest@Flatten[Permutations/@Subsets[Range@9,9],1],And@@IntegerQ/@(#/IntegerDigits@#)&]</lang>
- Output:
9867312
Nanoquery
Base 10
<lang nanoquery>s = ""
def uniqueDigits(i)
global s
// returns true, if unique digits, false otherwise for k in range(0, len(s) - 2) for l in range(k + 1, len(s) - 1) if (s[l] = "0") || (s[l] = "5") //0 or 5 digit return false end
if s[k] = s[l] //non-unique digit return false end end end
return true
end
def testNumber(i)
global s
//Tests, if i is divisible by all its digits (0 is not a digit already) j = 0 divisible = true for ch in s j = int(ch) divisible = (i % j) = 0 if not divisible return false end end
return true
end
i = 98764321 isUnique = true canBeDivided = true
while i > 0
s = str(i) isUnique = uniqueDigits(i) if isUnique //Number has unique digits canBeDivided = testNumber(i) if canBeDivided println "Number found: " + i i = 0 end end i -= 1
end</lang>
Nim
Base 10
This version uses a combination of several algorithms, especially those of the C++ and the Go/Raku versions. But, to get the digits, instead of converting the number to a string it uses an iterator which is significantly faster. And to check digit uniqueness, it uses a very efficient bit set. The programs runs in less than 20ms.
<lang nim>type Digit = range[0..9]
iterator digits(n: int64): Digit =
var n = n while true: yield n mod 10 n = n div 10 if n == 0: break
func isLynchBell(num: int64): bool =
var hSet: set[Digit] for d in num.digits: if d == 0 or num mod d != 0 or d in hSet: return false hSet.incl(d) return true
const
Magic = 9 * 8 * 7 High = 0x9876432 div Magic * Magic
for n in countdown(High, Magic, Magic):
if n.isLynchBell: echo n break</lang>
- Output:
9867312
Base 16
This is the same algorithm adapted for base 16. The program runs in about 30ms.
<lang Nim>import strformat
type Digit = range[0..15]
iterator digits(n: int64): Digit =
var n = n while true: yield n and 15 n = n shr 4 if n == 0: break
func isLynchBell(num: int64): bool =
var hSet: set[Digit] for d in num.digits: if d == 0 or num mod d != 0 or d in hSet: return false hSet.incl(d) return true
const Magic = 15 * 14 * 13 * 12 * 11 const High = 0xfedcba987654321 div Magic * Magic
for n in countdown(High, Magic, Magic):
if n.isLynchBell: echo &"{n:x}" break</lang>
- Output:
fedcb59726a1348
Perl
Base 10
<lang perl>my $step = 9 * 8 * 7; # 504, interval between tests
my $initial = int(9876432 / $step) * $step; # largest 7 digit multiple of 504 < 9876432
for($test = $initial; $test > 0 ; $test -= $step) { # decrement by 504
next if $test =~ /[05]/; # skip numbers containing 0 or 5 next if $test =~ /(.).*\1/; # skip numbers with non unique digits
for (split , $test) { # skip numbers that don't divide evenly by all digits next unless ($test / $_) % 1; }
printf "Found $test after %d steps\n", ($initial-$test)/$step; for (split , $test) { printf "%s / %s = %s\n", $test, $_, $test / $_; } last
}</lang>
- Output:
Found 9867312 after 18 steps 9867312 / 9 = 1096368 9867312 / 8 = 1233414 9867312 / 6 = 1644552 9867312 / 7 = 1409616 9867312 / 3 = 3289104 9867312 / 1 = 9867312 9867312 / 2 = 4933656
Base 16
<lang perl>use bigint; # Very slow, but consistent results even with 32-bit Perl
my $hex = 'FEDCBA987654321'; # largest possible hex number $step = Math::BigInt::blcm(1..15); $initial = int(hex($hex) / $step) * $step;
for($num = $initial; $num > 0 ; $num -= $step) { # decrement by lcm
my $test = sprintf '%x', $num; next if $test =~ /0/; # skip numbers containing 0 next if $test =~ /(.).*\1/; # skip numbers with non unique digits
push @res, sprintf "Found $test after %d steps\n", ($initial-$num)/$step; push @res, ' 'x12 . 'In base 16' . ' 'x36 . 'In base 10'; for (split , $test) { push @res, sprintf "%s / %s = %x | %d / %2d = %19d", $test, $_, $num / hex($_), $num, hex($_), $num / hex($_); } last
}
print join "\n", @res;</lang>
- Output:
Found fedcb59726a1348 after 954460 steps In base 16 In base 10 fedcb59726a1348 / f = 10fda5b4be4f038 | 1147797065081426760 / 15 = 76519804338761784 fedcb59726a1348 / e = 1234561d150b83c | 1147797065081426760 / 14 = 81985504648673340 fedcb59726a1348 / d = 139ad2e43e0c668 | 1147797065081426760 / 13 = 88292081929340520 fedcb59726a1348 / c = 153d0f21ede2c46 | 1147797065081426760 / 12 = 95649755423452230 fedcb59726a1348 / b = 172b56538f25ed8 | 1147797065081426760 / 11 = 104345187734675160 fedcb59726a1348 / 5 = 32f8f11e3aed0a8 | 1147797065081426760 / 5 = 229559413016285352 fedcb59726a1348 / 9 = 1c5169829283b08 | 1147797065081426760 / 9 = 127533007231269640 fedcb59726a1348 / 7 = 2468ac3a2a17078 | 1147797065081426760 / 7 = 163971009297346680 fedcb59726a1348 / 2 = 7f6e5acb93509a4 | 1147797065081426760 / 2 = 573898532540713380 fedcb59726a1348 / 6 = 2a7a1e43dbc588c | 1147797065081426760 / 6 = 191299510846904460 fedcb59726a1348 / a = 197c788f1d76854 | 1147797065081426760 / 10 = 114779706508142676 fedcb59726a1348 / 1 = fedcb59726a1348 | 1147797065081426760 / 1 = 1147797065081426760 fedcb59726a1348 / 3 = 54f43c87b78b118 | 1147797065081426760 / 3 = 382599021693808920 fedcb59726a1348 / 4 = 3fb72d65c9a84d2 | 1147797065081426760 / 4 = 286949266270356690 fedcb59726a1348 / 8 = 1fdb96b2e4d4269 | 1147797065081426760 / 8 = 143474633135178345
Phix
base 10
with javascript_semantics integer magic = 9*8*7, high = 9876432, n = high-mod(high,magic) sequence seen while true do string s = sprintf("%d",n) seen = {1,0,0,0,0,1,0,0,0,0} for j=1 to length(s) do seen[s[j]-'0'+1] += 1 end for if max(seen)=1 then exit end if n -= magic end while -- may as well quickly verify... seen[5+1] = 0 -- (skipping 5) for i=1 to 9 do -- ( and 0 ) if seen[i+1] then if mod(n,i)!=0 then ?9/0 end if end if end for printf(1,"%d (%d iterations)\n",{n,(high-n)/magic})
- Output:
9867312 (18 iterations)
base 16
with javascript_semantics atom t0 = time() integer count = 0 string s include mpfr.e mpz lcm15 = mpz_init(lcm(tagset(15))), d = mpz_init(), r = mpz_init() mpz_set_str(d,"FEDCBA987654321",16) mpz_mod(r,d,lcm15) mpz_sub(d,d,r) -- d:= max k*lcm <= "FE..21" r = mpz_free(r) -- (no longer used) while true do s = mpz_get_str(d,16) if sort(s)="123456789abcdef" then exit end if mpz_sub(d,d,lcm15) count += 1 end while string e = elapsed(time()-t0) printf(1,"%s (%d iterations, %s)\n",{s,count,e})
- Output:
fedcb59726a1348 (954460 iterations, 12.4s)
Prolog
This will work with any radix, including base 10 and base 16. <lang prolog>% Find the largest integer divisible by all it's digits, with no digit repeated. % ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ % We go for a generic algorithm here. Test for divisibility is done by % calculating the least common multiplier for all digits, and testing % whether a candidate can be divided by the LCM without remainder. % % Instead of iterating numbers and checking whether the number has % repeating digits, it is more efficient to generate permutations of % digits and then convert to a number. Doing it this way reduces search % space greatly. % % Notes: % For decimal numbers we could improve times by testing only numbers % of length 7 (since 5x2=10 and 0 is not one of our digits, and 9x2=18 % which needs 2 digits to store), but that sort of logic does not % hold for hexadecimal numbers. % We could also explicitly eliminate odd numbers, but the double validity % check actually slows us down very slightly instead of speeding us up.
- - dynamic
trial/1. % temporarily store digit combinations here.
gcd(X, X, X). % Calculate greatest common divisor gcd(M, N, X) :- N > M, B is N-M, gcd(M,B,X). gcd(M, N, X) :- N < M, A is M-N, gcd(A,N,X).
lcm(A, B, LCM) :- gcd(A,B,GCD), LCM is A * B / GCD.
lcm([H], H). % Calculate least common multiplier lcm([A|T], LCM) :- lcm(T, B), !, lcm(A,B,LCM).
mkint(_, Val, [], Val). % Result = Val where list is empty mkint(Radix, Val, [H|T], Int) :- % (((I0*10+I1)*10+I2)*10+In)... V0 is Val*Radix+H, !, mkint(Radix, V0, T, Int).
% Turn a list of digits into an integer number using Radix. mkint(Radix, [H|T], Int) :- mkint(Radix, H, T, Int).
domain(0, []). % For example, domain(5) is [1,2,3,4,5] domain(N, [N|Digits]) :-
succ(N0, N), !, domain(N0, Digits).
trial(0, Digits, Digits). % generates a combination of digits to test trial(N, D, Digits) :- % remove N digits, and find remaining combinations
append(L0,[_|L1],D), succ(N0, N), trial(N0, L1, Dx), append(L0, Dx, Digits). % trial(1, [3,2,1], D) -> D=[2,1]; D=[3,1]; D=[3,2].
make_trials(_,_) :- retractall(trial(_)), fail. make_trials(N,Domain) :- trial(N, Domain, Digits), asserta(trial(Digits)), fail. make_trials(_,_). % trials are stored highest values to lowest
combinations(Radix, NDigits) :- % Precalculate all possible digit combinations
succ(R0, Radix), domain(R0, Domain), Nskip is R0 - NDigits, make_trials(Nskip, Domain).
test(Radix, Digits, LCM, Number) :- % Make an integer and check for divisibility
mkint(Radix, Digits, Number), 0 is Number mod LCM.
bignum(Radix, Number) :-
succ(R0, Radix), between(1,R0,N), NDigits is Radix - N, % loop decreasing length combinations(Radix, NDigits), % precalc digit combos with length=NDigits trial(Digits), lcm(Digits, LCM), % for a combination, calculate LCM permutation(Digits, Trial), % generate a permutation test(Radix, Trial, LCM, Number). % test for divisibility
largest_decimal(N) :- bignum(10, N), !. largest_hex(N, H) :- bignum(16, N), !, sformat(H, '~16r', [N]).</lang>
?- time(largest_decimal(S)). % 20,043,250 inferences, 3.086 CPU in 3.089 seconds (100% CPU, 6493905 Lips) S = 9867312. ?- time(largest_hex(S,H)). % 73,332,059 inferences, 11.800 CPU in 11.803 seconds (100% CPU, 6214553 Lips) S = 1147797065081426760, H = "fedcb59726a1348".
Python
base 10
Using the insights presented in the preamble to the Raku (base 10) example:
<lang python>Largest number divisible by its digits
from itertools import (chain, permutations) from functools import (reduce) from math import (gcd)
- main :: IO ()
def main():
Tests
# (Division by zero is not an option, so 0 and 5 are omitted) digits = [1, 2, 3, 4, 6, 7, 8, 9]
# Least common multiple of the digits above lcmDigits = reduce(lcm, digits)
# Any 7 items drawn from the digits above, # including any two of [1, 4, 7] sevenDigits = ((delete)(digits)(x) for x in [1, 4, 7])
print( max( ( intFromDigits(x) for x in concatMap(permutations)(sevenDigits) ), key=lambda n: n if 0 == n % lcmDigits else 0 ) )
- intFromDigits :: [Int] -> Int
def intFromDigits(xs):
An integer derived from an ordered list of digits. return reduce(lambda a, x: a * 10 + x, xs, 0)
- ----------------------- GENERIC ------------------------
- concatMap :: (a -> [b]) -> [a] -> [b]
def concatMap(f):
A concatenated list over which a function has been mapped. The list monad can be derived by using a function f which wraps its output in a list, (using an empty list to represent computational failure). def go(xs): return chain.from_iterable(map(f, xs)) return go
- delete :: Eq a => [a] -> a -> [a]
def delete(xs):
xs with the first instance of x removed. def go(x): ys = xs.copy() ys.remove(x) return ys return go
- lcm :: Int -> Int -> Int
def lcm(x, y):
The smallest positive integer divisible without remainder by both x and y. return 0 if (0 == x or 0 == y) else abs( y * (x // gcd(x, y)) )
- MAIN ---
if __name__ == '__main__':
main()</lang>
- Output:
9867312
base 16
Descending from the upper limit, in steps of 360360 (least common multiple of the fifteen digit values), until the first number that uses all fifteen digits when expressed in hexadecimal.
<lang python>Largest number divisible by its hex digits
from functools import (reduce) from math import (gcd)
- main :: IO ()
def main():
First integer evenly divisible by each of its hex digits, none of which appear more than once.
# Least common multiple of digits [1..15] # ( -> 360360 ) lcmDigits = foldl1(lcm)( enumFromTo(1)(15) ) allDigits = 0xfedcba987654321
# ( -> 1147797409030632360 ) upperLimit = allDigits - (allDigits % lcmDigits)
# Possible numbers xs = enumFromThenToNext(upperLimit)( upperLimit - lcmDigits )(1)
print( hex( until(lambda x: 15 == len(set(showHex(x))))( lambda _: next(xs) )(next(xs)) ) ) # --> 0xfedcb59726a1348
- ------------------ GENERIC FUNCTIONS -------------------
- enumFromThenToNext :: Int -> Int -> Int -> Gen [Int]
def enumFromThenToNext(m):
Non-finite series of integer values enumerated from m to n with a step size defined by nxt-m. def go(m, nxt): d = nxt - m v = m while True: yield v v = d + v return lambda nxt: lambda n: go(m, nxt)
- enumFromTo :: (Int, Int) -> [Int]
def enumFromTo(m):
Integer enumeration from m to n. return lambda n: range(m, 1 + n)
- foldl1 :: (a -> a -> a) -> [a] -> a
def foldl1(f):
Left to right reduction of the non-empty list xs, using the binary operator f, with the head of xs as the initial acccumulator value. return lambda xs: reduce( lambda a, x: f(a)(x), xs[1:], xs[0] ) if xs else None
- lcm :: Int -> Int -> Int
def lcm(x):
The smallest positive integer divisible without remainder by both x and y. return lambda y: ( 0 if (0 == x or 0 == y) else abs( y * (x // gcd(x, y)) ) )
- until :: (a -> Bool) -> (a -> a) -> a -> a
def until(p):
The result of repeatedly applying f until p holds. The initial seed value is x. def go(f, x): v = x while not p(v): v = f(v) return v return lambda f: lambda x: go(f, x)
- showHex :: Int -> String
def showHex(n):
Hexadecimal string representation of an integer value. return hex(n)[2:]
- MAIN --
if __name__ == '__main__':
main()</lang>
- Output:
0xfedcb59726a1348 [Finished in 1.243s]
Raku
(formerly Perl 6)
Base 10
The number can not have a zero in it, that implies that it can not have a 5 either since if it has a 5, it must be divisible by 5, but the only numbers divisible by 5 end in 5 or 0. It can't be zero, and if it is odd, it can't be divisible by 2, 4, 6 or 8. So that leaves 98764321 as possible digits the number can contain. The sum of those 8 digits is not divisible by three so the largest possible integer must use no more than 7 of them (since 3, 6 and 9 would be eliminated). Strictly by removing possibilities that cannot possibly work we are down to at most 7 digits.
We can deduce that the digit that won't get used is one of 1, 4, or 7 since those are the only ones where the removal will yield a sum divisible by 3. It is extremely unlikely be 1, since EVERY number is divisible by 1. Removing it reduces the number of digits available but doesn't gain anything as far as divisibility. It is unlikely to be 7 since 7 is prime and can't be made up of multiples of other numbers. Practically though, the code to accommodate these observations is longer running and more complex than just brute-forcing it from here.
In order to accommodate the most possible digits, the number must be divisible by 7, 8 and 9. If that is true then it is automatically divisible by 2, 3, 4, & 6 as they can all be made from the combinations of multiples of 2 and 3 which are present in 8 & 9; so we'll only bother to check multiples of 9 * 8 * 7 or 504.
All these optimizations get the run time to well under 1 second.
<lang perl6>my $magic-number = 9 * 8 * 7; # 504
my $div = 9876432 div $magic-number * $magic-number; # largest 7 digit multiple of 504 < 9876432
for $div, { $_ - $magic-number } ... * -> $test { # only generate multiples of 504
next if $test ~~ / <[05]> /; # skip numbers containing 0 or 5 next if $test ~~ / (.).*$0 /; # skip numbers with non unique digits
say "Found $test"; # Found a solution, display it for $test.comb { printf "%s / %s = %s\n", $test, $_, $test / $_; } last
}</lang>
- Output:
Found 9867312 9867312 / 9 = 1096368 9867312 / 8 = 1233414 9867312 / 6 = 1644552 9867312 / 7 = 1409616 9867312 / 3 = 3289104 9867312 / 1 = 9867312 9867312 / 2 = 4933656
Base 16
There are fewer analytical optimizations available for base 16. Other than 0, no digits can be ruled out so a much larger space must be searched. We'll start at the largest possible permutation (FEDCBA987654321) and work down so as soon as we find a solution, we know it is the solution. The combination of .race
with .first
lets us utilize concurrency and exit early when the single desired solution is found.
<lang perl6>my $hex = 'FEDCBA987654321'; # largest possible hex number my $magic-number = [lcm] 1 .. 15; # find least common multiple my $div = :16($hex) div $magic-number * $magic-number;
- hunt for target stepping backwards in multiples of the lcm
my $target = ($div, * - $magic-number ... 0).race.first: -> \test {
my \num= test.base(16); (num.contains('0') || num.comb.Bag.values.max > 1) ?? False !! True
}; my $hexnum = $target.base(16);
say "Found $hexnum"; # Found a solution, display it
say ' ' x 12, 'In base 16', ' ' x 36, 'In base 10'; for $hexnum.comb {
printf "%s / %s = %s | %d / %2d = %19d\n", $hexnum, $_, ($target / :16($_)).base(16), $target, :16($_), $target / :16($_);}</lang>
- Output:
Found FEDCB59726A1348 In base 16 In base 10 FEDCB59726A1348 / F = 10FDA5B4BE4F038 | 1147797065081426760 / 15 = 76519804338761784 FEDCB59726A1348 / E = 1234561D150B83C | 1147797065081426760 / 14 = 81985504648673340 FEDCB59726A1348 / D = 139AD2E43E0C668 | 1147797065081426760 / 13 = 88292081929340520 FEDCB59726A1348 / C = 153D0F21EDE2C46 | 1147797065081426760 / 12 = 95649755423452230 FEDCB59726A1348 / B = 172B56538F25ED8 | 1147797065081426760 / 11 = 104345187734675160 FEDCB59726A1348 / 5 = 32F8F11E3AED0A8 | 1147797065081426760 / 5 = 229559413016285352 FEDCB59726A1348 / 9 = 1C5169829283B08 | 1147797065081426760 / 9 = 127533007231269640 FEDCB59726A1348 / 7 = 2468AC3A2A17078 | 1147797065081426760 / 7 = 163971009297346680 FEDCB59726A1348 / 2 = 7F6E5ACB93509A4 | 1147797065081426760 / 2 = 573898532540713380 FEDCB59726A1348 / 6 = 2A7A1E43DBC588C | 1147797065081426760 / 6 = 191299510846904460 FEDCB59726A1348 / A = 197C788F1D76854 | 1147797065081426760 / 10 = 114779706508142676 FEDCB59726A1348 / 1 = FEDCB59726A1348 | 1147797065081426760 / 1 = 1147797065081426760 FEDCB59726A1348 / 3 = 54F43C87B78B118 | 1147797065081426760 / 3 = 382599021693808920 FEDCB59726A1348 / 4 = 3FB72D65C9A84D2 | 1147797065081426760 / 4 = 286949266270356690 FEDCB59726A1348 / 8 = 1FDB96B2E4D4269 | 1147797065081426760 / 8 = 143474633135178345
Red
<lang Rebol>Red [] t0: now/time/precise ;; measure runtime lbn: 98764321 + 1 ;; because digit 5 is ruled out, this is the highest 8 digit number
;; possible, add 1 because only even numbers are possible
check: func [tos [ string! ]] [ ;; function to check if number is divideable by
foreach ele tos [ ;; all of its digits div: to-integer ele - #"0" ;; convert asci digit to integer unless lbn % div = 0 [ return false ] ;; fail at first false condition ( unless = if not...) ] true ;; true if all digits passed
]
forever [
lbn: lbn - 2 ;; only even numbers could be possible results if find tos: to-string lbn "0" [continue] ;; no "0" allowed if find tos "5" [continue] ;; "5" also excluded unless tos = unique tos [ continue ] ;; only unique digits allowed unless check tos [continue] ;; passed check ? print lbn ;; first hit is result probe now/time/precise - t0 ;; display runtime halt
] </lang>
- Output:
(interpreted version
9867312
0:01:38.004 (halted)
REXX
base 10
This REXX version uses mostly the same logic and deductions that the Raku example does, but it performs
the tests in a different order for maximum speed.
The inner do loop is only executed a score of times; the 1st if statement does the bulk of the eliminations. <lang rexx>/*REXX program finds the largest (decimal) integer divisible by all its decimal digits. */ $= 7 * 8 * 9 /*a # that it must divide the found #. */ t= 0 /*the number of divisibility trials. */
do #=9876432 % $ * $ by -$ /*search from largest number downwards.*/ if # // $ \==0 then iterate /*Not divisible? Then keep searching.*/ if verify(50, #, 'M') \==0 then iterate /*does it contain a five or a zero? */ t= t+1 /*curiosity's sake, track # of trials. */ do j=1 for length(#) - 1 /*look for a possible duplicated digit.*/ if pos( substr( #, j, 1), #, j+1) \==0 then iterate # end /*j*/ /* [↑] Not unique? Then keep searching*/ /* [↓] superfluous, but check anyways.*/ do v=1 for length(#) /*verify the # is divisible by all digs*/ if # // substr(#, v, 1) \==0 then iterate # end /*v*/ /* [↑] ¬divisible? Then keep looking.*/ leave /*we found a number, so go display it. */ end /*#*/
say 'found ' # " (in " t ' trials)' /*stick a fork in it, we're all done. */</lang>
- output:
Timing note: execution time is under 1/2 millisecond (essentially not measurable in the granularity of the REXX timer under Microsoft Windows).
found 9867312 (in 11 trials)
base 16
The "magic" number was expanded to handle hexadecimal numbers.
Note that 15×14×13×12×11 is the same as 13×11×9×8×7×5. <lang rexx>/*REXX program finds the largest hexadecimal integer divisible by all its hex digits. */ numeric digits 20 /*be able to handle the large hex nums.*/ bigH= 'fedcba987654321' /*biggest number possible, hexadecimal.*/ bigN= x2d(bigH) /* " " " decimal. */ $= 15 * 14 * 13 * 12 * 11 /*a # that it must divide the found #. */ t= 0 /*the number of divisibility trials. */
do #=bigN % $ * $ by -$ /*search from largest poss. # downwards*/ if # // $ \==0 then iterate /*Not divisible? Then keep searching.*/ h= d2x(#) /*convert decimal number to hexadecimal*/ if pos(0, h) \==0 then iterate /*does hexadecimal number contain a 0? */ t= t+1 /*curiosity's sake, track # of trials. */ do j=1 for length(h) - 1 /*look for a possible duplicated digit.*/ if pos( substr(h, j, 1), h, j+1) \==0 then iterate # end /*j*/ /* [↑] Not unique? Then keep searching*/
do v=1 for length(h) /*verify the # is divisible by all digs*/ if # // x2d(substr( h, v, 1) ) \==0 then iterate # end /*v*/ /* [↑] ¬divisible? Then keep looking.*/ leave /*we found a number, so go display it. */ end /*#*/
say 'found ' h " (in " t ' trials)' /*stick a fork in it, we're all done. */</lang>
- output:
found FEDCB59726A1348 (in 287747 trials)
Ring
<lang ring>
- Project : Largest number divisible by its digits
for n = 9867000 to 9867400
numbers = list(9) for t=1 to 9 numbers[t] = 0 next flag = 1 flag2 = 1 flag3 = 1 str=string(n) for m=1 to len(str) if number(str[m]) > 0 numbers[number(str[m])] = numbers[number(str[m])] + 1 else flag2 = 0 ok next if flag2 = 1 for p=1 to 9 if numbers[p] = 0 or numbers[p] = 1 else flag = 0 ok next if flag = 1 for x=1 to len(str) if n%(number(str[x])) != 0 flag3 = 0 ok next if flag3 = 1 see n + nl ok ok ok
next </lang> Output:
9867312
Ruby
base 10
Following the reasoning of the Raku sample. <lang ruby>magic_number = 9*8*7 div = 9876432.div(magic_number) * magic_number candidates = div.step(0, -magic_number)
res = candidates.find do |c|
digits = c.digits (digits & [0,5]).empty? && digits == digits.uniq
end
puts "Largest decimal number is #{res}"</lang>
- Output:
Largest decimal number is 9867312
base 16
<lang ruby>def divByAll(num, digits)
digits.all? { |digit| num % digit.to_i(16) == 0 }
end
magic = 15 * 14 * 13 * 12 * 11 high = (0xfedcba987654321 / magic) * magic
high.step(magic, -magic) do |i|
s = i.to_s(16) # always generates lower case a-f next if s.include? "0" # can't contain '0' sd = s.chars.uniq next if sd.size != s.size # digits must be unique (puts "Largest hex number is #{i.to_s(16)}"; break) if divByAll(i, sd)
end</lang>
- Output:
Largest hex number is fedcb59726a1348
Scala
base 10
This example starts with a lazily evaluated list of decreasing decimal numbers, starting with 98764321 (5 is eliminated as per the Pearl 6 analysis). It applies a filter to only accept numbers with distinct, nonzero digits that all divide the number itself, and then returns the head of the list.
<lang scala>import scala.collection.mutable
def largestDecimal: Int = Iterator.from(98764321, -1).filter(chkDec).next def chkDec(num: Int): Boolean = {
val set = mutable.HashSet[Int]() num.toString.toVector.map(_.asDigit).forall(d => (d != 0) && (num%d == 0) && set.add(d))
}</lang>
- Output:
scala> println(s"Base 10: $largestDecimal") Base 10: 9867312
base 16
While concise, the previous example is relatively slow, taking nearly 30 seconds to complete. So, instead of simply moving on to a base 16 version, this next example is a fast version for arbitrary base. Starting with a list of digits generated from the given base, the program generates a lazily evaluated list of all possible combinations of digits in blocks of decreasing length. Each block is passed to a function that generates a list of numbers for each combination which are divisible by all the digits, then filters it for numbers which are made up of the required digits. The blocks are checked in order until a number is found.
<lang scala>import spire.math.SafeLong import spire.implicits._
object LargestNumDivisibleByDigits {
def main(args: Array[String]): Unit = { for(b <- Seq(10, 16)){ val tStart = System.currentTimeMillis val res = getLargestNum(b).toBigInt.toString(b) val tDur = System.currentTimeMillis - tStart println(s"Base $b: $res [${tDur}ms]") } } def getLargestNum(base: SafeLong): SafeLong = { def chkNum(digits: Vector[SafeLong])(num: SafeLong): Boolean = { val lst = LazyList.iterate((num%base, num/base)){case (_, src) => (src%base, src/base)}.take(digits.length).map(_._1) lst.diff(digits).isEmpty } def chkChunk(combo: Vector[SafeLong]): Option[SafeLong] = { val lcm = combo.reduce(_.lcm(_)) val ulim = combo.zipWithIndex.map{case (n, i) => n*(base ** i)}.reduce(_+_) Iterator.iterate(ulim - (ulim%lcm))(_ - lcm).takeWhile(_ > 0).find(chkNum(combo)) } val baseDigits: Vector[SafeLong] = Vector.range(1, base.toInt).map(SafeLong(_)) def chkBlock(digits: Iterator[Vector[SafeLong]]): Option[SafeLong] = digits.map(chkChunk).collect{case Some(n) => n}.maxOption Iterator.from(base.toInt - 1, -1).map(len => chkBlock(baseDigits.combinations(len))).collect{case Some(n) => n}.next }
}</lang>
- Output:
Base 10: 9867312 [1144ms] Base 16: fedcb59726a1348 [1090ms]
Sidef
base 10
<lang ruby>func largest_number(base) {
var digits = @(base ^.. 1)
digits.each {|k| digits.variations(k, {|*a| var n = Number(a.join, base) if (a.all {|d| d.divides(n) }) { return n } }) }
}
say largest_number(10) #=> 9867312</lang>
VBScript
base 10
<lang vb>' Largest number divisible by its digits - base10 - VBScript s=7*8*9 'reasonable assumption m=9876432 'reasonable assumption for i=(m\s)*s to 1 step -s if instr(i,"5")=0 and instr(i,"0")=0 then '5 or 0 impossible b=false: j=1 while j<=len(i)-1 and not b if instr(j+1,i,mid(i,j,1))<>0 then b=true 'no duplicated digit j=j+1 wend if not b then j=1 while j<=len(i) and not b if (i mod mid(i,j,1))<>0 then b=true 'divisible by all digits j=j+1 wend if not b then exit for end if end if next wscript.echo i </lang>
- Output:
9867312
Visual Basic .NET
<lang vbnet>Module Module1
Function ChkDec(num As Integer) As Boolean Dim sett As New HashSet(Of Integer) Return num.ToString() _ .Select(Function(c) Asc(c) - Asc("0")) _ .All(Function(d) (d <> 0) AndAlso (num Mod d = 0) AndAlso sett.Add(d)) End Function
Sub Main() Dim result = Enumerable.Range(0, 98764321) _ .Reverse() _ .Where(AddressOf ChkDec) _ .First() Console.WriteLine(result) End Sub
End Module</lang>
- Output:
9867312
Wren
base 10
<lang ecmascript>var divByAll = Fn.new { |n, digits| digits.all { |d| n%(d-48) == 0 } }
var magic = 9 * 8 * 7 var high = (9876432/magic).floor * magic var i = high while (i >= magic) {
if (i%10 != 0) { // can't end in '0' var s = "%(i)" if (!s.contains("0") && !s.contains("5")) { // can't contain '0' or '5' var set = {} var sd = [] // list of distinct digits for (b in s.bytes) { if (set[b] == null) { set[b] = true sd.add(b) } } if (sd.count == s.count) { // digits must be unique if (divByAll.call(i, sd)) { System.print("Largest decimal number is %(i)") return } } } } i = i - magic
}</lang>
- Output:
Largest decimal number is 9867312
base 16
The integers here are too large (>= 2^53) to be accurately represented by Wren and so we follow the AWK approach of using an array of digits to represent them. <lang ecmascript>var digits = "0123456789abcdef" var base = 16 var size = 15 var comDiv = 15 * 14 * 13 * 12 * 11
// Returns n mod k, where n is an array and k is a number var hexMod = Fn.new { |n, k|
var r = 0 for (i in size..1) r = (r*base + n[i]) % k return r
}
// Calculates n = n - m, where n is an array and m is a number var hexSub = Fn.new { |n, m|
var i = 1 while (m != 0 && i <= size) { n[i] = n[i] - (m%base) m = (m/base).floor if (n[i] < 0) { n[i] = n[i] + base m = m + 1 } i = i + 1 }
}
// Returns true if n is an array of unique digits in range 1..(base-1) var hasUniqueDigits = Fn.new { |n|
var dcount = List.filled(base, 0) for (i in 1..size) { var d = n[i] if (d == 0) return false // can't contain '0' dcount[d] = dcount[d] + 1 if (dcount[d] > 1) return false // digits must be unique } return true
}
var solve = Fn.new { |n|
var r = hexMod.call(n, comDiv) hexSub.call(n, r) while (n[size] > 0) { if (hasUniqueDigits.call(n)) { System.write("Largest hex number is ") for (i in size..1) System.write(digits[n[i]]) System.print() return } hexSub.call(n, comDiv) }
}
var startN = List.filled(size + 1, 0) for (i in 1..size) startN[i] = i solve.call(startN)</lang>
- Output:
Largest hex number is fedcb59726a1348
zkl
base 10
<lang zkl>const magic_number=9*8*7; # 504 const div=9876432 / magic_number * magic_number; #largest 7 digit multiple of 504 < 9876432
foreach test in ([div..0,-magic_number]){
text:=test.toString(); if(text.holds("0","5")) continue; # skip numbers containing 0 or 5 if(text.unique().len()!=text.len()) continue; # skip numbers with non unique digits if(test.split().filter1('%.fp(test))) continue; # skip numbers that don't divide evenly by all digits println("Found ",test); # Found a solution, display it foreach d in (test.split()){ println("%s / %s = %s".fmt(test,d, test/d)); } break;
}</lang>
- Output:
Found 9867312 9867312 / 9 = 1096368 9867312 / 8 = 1233414 9867312 / 6 = 1644552 9867312 / 7 = 1409616 9867312 / 3 = 3289104 9867312 / 1 = 9867312 9867312 / 2 = 4933656
base 16
<lang zkl>const bigN=0xfedcba987654321; // biggest hexadecimal number possible. lcm:=lcmNs([1..15]); // 360360, smallest # that will divide answer upperLimit:=bigN - bigN%lcm; // start at a mulitple of whatever the answer is
foreach test in ([upperLimit..1,-lcm]){
text:=test.toString(16); if(15!=text.unique().len()) continue; println(text); break;
}</lang> <lang zkl>fcn lcmNs(ns){ ns.reduce(fcn(m,n){ (m*n).abs()/m.gcd(n) }) }</lang>
- Output:
fedcb59726a1348