Sub-unit squares
You are encouraged to solve this task according to the task description, using any language you may know.
A sub-unit square is a square number (product of two identical non-negative integers) that remains a square after having a 1 subtracted from each digit in the square.
- E.G.
The number 1 is a sub-unit square. 1 - 1 is 0, which is also a square, though it's kind-of a degenerate case.
The number 3136 is a sub-unit square. 3136 (56²) with unit 1 subtracted from each digit is 2025 (45²).
A sub-unit square cannot contain a digit zero (0) since zero minus one is negative one. Every known sub-unit square, with the exception of 1, ends with the digits 36.
- Task
- Find and display at least the first five sub-unit squares.
- See also
ALGOL 68
Assumes LONG INT is at least 64 bit (as in ALGOL 68G version 2 or 3). Uses the optimisations in the J sample, based on Wherra's observation that the squares (other than 1) must end in 36 (see the Discussion page).
BEGIN # find some sub-unit squares: squares which are still squares when 1 #
# is subtracted from each of their digits #
# the sub-unit squares must end in 36 (see task discussion) and so #
# the square formed by subtracting 1 from each digit must end in 25 #
# so, as with the J sample, we start with the "25" square and test #
# the incremented square #
# find the sub unit squares - note that 1 is a special case #
print( ( "1" ) );
INT su count := 1;
LONG INT add ones := 1;
LONG INT power of 10 := 10;
FOR n FROM 5 BY 10 WHILE su count < 7 DO
LONG INT sq = ( LENG n ) * ( LENG n );
WHILE sq > power of 10 DO
# the square has more digits than the previous one #
power of 10 *:= 10;
add ones *:= 10 +:= 1
OD;
IF LONG INT add unit sq = sq + add ones;
add unit sq < power of 10
THEN
# squaring the number with one added to the digits has the #
# same number of digits as the original square #
IF LONG INT root = ENTIER long sqrt( add unit sq );
root * root = add unit sq
THEN
# the add unit square is actually a square #
# make sure there are no 0s in the add unit square as they #
# can't be decremented to a positive digit #
IF BOOL no zeros := TRUE;
LONG INT v := add unit sq;
WHILE v > 0 AND no zeros DO
no zeros := v MOD 10 /= 0;
v OVERAB 10
OD;
no zeros
THEN
print( ( " ", whole( add unit sq, 0 ) ) );
su count +:= 1
FI
FI
FI
OD
END
- Output:
1 36 3136 24336 5973136 71526293136 318723477136
Arturo
allSquares: map 1..1000000 'x -> x^2
subUnits: allSquares | select 'z -> not? some? digits z 'w -> w=0
| select 'x -> contains? allSquares to :integer join map digits x 'k -> k-1
print ["The first" size subUnits "sub-unit squares:"]
print subUnits
- Output:
The first 10 sub-unit squares: 36 3136 24336 118336 126736 5973136 71526293136 113531259136 137756776336 318723477136
Delphi
function IsSubUnitSquare(SQ: integer): boolean;
{Returns true if you subtract one from each digit}
{and it is still square. Assume SQ is square}
var I,J: integer;
var IA: TIntegerDynArray;
begin
Result:=False;
{Get all digits}
GetDigits(SQ,IA);
{Subtract one from each digit}
for J:=0 to High(IA) do
begin
{Zeros not allowed = they would cause negative digits}
if IA[J]=0 then exit;
IA[J]:=IA[J]-1;
end;
{Turn digits into number again}
SQ:=DigitsToInt(IA);
{Test if it is square}
if Frac(Sqrt(SQ))<>0 then exit;
Result:=True;
end;
procedure ShowSubUnitSquares(Memo: TMemo);
var I,SQ,Cnt: integer;
begin
Cnt:=0;
for I:=1 to high(Integer) do
begin
SQ:=I*I;
if IsSubUnitSquare(SQ) then
begin
Inc(Cnt);
Memo.Lines.Add(Format('%d %8.0n',[Cnt,SQ+0.0]));
if Cnt>=7 then break;
end;
end;
end;
- Output:
1 1 2 36 3 3,136 4 24,336 5 118,336 6 126,736 7 5,973,136 Elapsed Time: 6.899 ms.
EasyLang
func has9 v .
while v > 0
if v mod 10 = 9
return 1
.
v = v div 10
.
return 0
.
ones = 1
pow10 = 10
while count < 5
while n * n > pow10
pow10 *= 10
ones = ones * 10 + 1
.
if has9 (n * n) = 0
sq = n * n + ones
if sqrt sq mod 1 = 0
write " " & sq
count += 1
.
.
n += 1
.
FreeBASIC
Function HasZero(n As Uinteger) As Boolean
Return Iif(Instr(Str(n), "0"), True, False)
End Function
Function DecDigits(n As Uinteger) As Uinteger
Dim As Integer t, m = 1111111111
Do
t = n - m
m /= 10
Loop Until t >= 0
Return t
End Function
Dim As Uinteger n = 1, c = 0
Dim As Uinteger n2, m, m2
Do
n2 = n*n
If Not HasZero(n2) Then
m2 = DecDigits(n2)
m = Sqr(m2)
If m*m = m2 Then
Print n2 & " ";
c += 1
End If
End If
n += 1
Loop Until c >= 7
Sleep
- Output:
Same as XPL0 entry.
Haskell
import Data.Char ( digitToInt , intToDigit )
isSquareNumber :: Int -> Bool
isSquareNumber n = root * root == n
where
root :: Int
root = floor $ sqrt $ fromIntegral n
isSubunitSquare :: Int -> Bool
isSubunitSquare n
|elem '0' numstr = False
|otherwise = isSquareNumber n && isSquareNumber subunitnum
where
numstr :: String
numstr = show n
subunitnum :: Int
subunitnum = read $ map (intToDigit . pred . digitToInt ) numstr
solution :: [Int]
solution = take 7 $ filter isSubunitSquare [1,2..]
- Output:
[1,36,3136,24336,118336,126736,5973136]
J
digits=: 10&#.inv
idigs=: >:&.digits"0
(#~ 1-(0 e. digits)"0) (#~ (=<.)@%:)idigs*:5*i.2e5
1 36 3136 24336 5973136 71526293136 318723477136
Here we work in the reverse direction. Since these sub-unit squares always end in 36, their decremented digit version must always end in 25 and the square roots must always end in 5. So, we start with a sequence of multiples of 5, square them, increment their digits and discard those which are not exact squares. Finally, we discard those which contain a 0 digit. (The method used to increment digits here maps 99 to 1010. Arguably, that should instead be 110, but it would get discarded, regardless. (With a little more complexity we could discard anything with a 9 digit while generating the initial sequence, this would provide only a minor gain.))
Java
import java.util.ArrayList;
import java.util.List;
public class Sub_unitSquares {
public static void main(String[] args) {
final int sub_unitsRequired = 7;
System.out.println("The first " + sub_unitsRequired + " sub-unit squares are:");
System.out.println(1);
long number = 2;
int count = 1;
while ( count < sub_unitsRequired ) {
final long square = number * number;
List<Integer> digits = digits(square);
if ( ! digits.contains(0) && digits.get(0) != 1 ) {
final long sub_unit = sub_unit(digits);
if ( isSquare(sub_unit) ) {
System.out.println(square);
count += 1;
}
}
number += 1;
}
}
private static boolean isSquare(long number) {
final long squareRoot =(long) Math.sqrt(number);
return number == squareRoot * squareRoot;
}
private static long sub_unit(List<Integer> digits) {
long result = 0;
for ( int digit : digits ) {
result = 10 * result + ( digit - 1 );
}
return result;
}
private static List<Integer> digits(long number) {
List<Integer> result = new ArrayList<Integer>();
while ( number != 0 ) {
result.add(0, (int) ( number % 10 ) );
number /= 10;
}
return result;
}
}
- Output:
The first 7 sub-unit squares are: 1 36 3136 24336 5973136 71526293136 318723477136
jq
The following jq program closely follows the task description and thus includes squares greater than 1 that begin with "1", even though A061844 excludes them.
Since gojq, the Go implementation of jq, uses unbounded-precision integer arithmetic, the program can be expected to produce reliable results when used with that implementation.
By contrast, the C implementation of jq uses IEEE 754 64-bit arithmetic, and so cannot be expected to produce reliable values beyond the first few, though in fact the results using the C-implementation are accurate through at least the first 11 sub-unit squares.
# The following is sufficiently accurate for the task:
def is_square: sqrt | . == floor;
# Emit a stream of "sub-unit squares"
def subunitSquares:
def sub1:
tostring | explode
| if any(. == 48) then null
else map(.-1) | implode | tonumber
end;
foreach range(1; infinite) as $i ( null;
($i * $i) as $sq
| ($sq|sub1)
| if . == null or (tostring[-1:] | IN("3", "7", "8")) then null
elif is_square then $sq
else null
end;
select(.) ) ;
11 as $count
| "The first \($count) sub-unit squares are:",
limit($count; subunitSquares)
- Output:
"The first 11 sub-unit squares are:" 1 36 3136 24336 118336 126736 5973136 71526293136 113531259136 137756776336 318723477136
Julia
""" Rosetta Code task rosettacode.org/wiki/Sub-unit_squares """
issquare(n::Integer) = isqrt(n)^2 == n
from_subsquare(n) = (d = digits(n); all( x -> x < 9, d) && issquare(evalpoly(10, d .+ 1)))
println("Subunit squares up to 10^12: ",
pushfirst!([evalpoly(10, digits(i^2) .+ 1) for i in 5:10:10^6 if from_subsquare(i^2)], 1))
- Output:
Subunit squares up to 10^12: [1, 36, 3136, 24336, 5973136, 71526293136, 318723477136]
Alternative procedural version.
# find some sub-unit squares: squares which are still squares when 1
# is subtracted from each of their digits
# the sub-unit squares must end in 36 (see task discussion) and so
# the square formed by subtracting 1 from each digit must end in 25
# so, as with the J sample, we start with the "25" square and test
# the incremented square
begin # find the sub unit squares - note that 1 is a special case
local maxCount = 7
local suSquares = collect( 1 : maxCount ) # array initially 1, 2, 3, ...
local suCount = 1
local addOnes = 1
local powerOf10 = 10
local n = -5
while suCount < maxCount
n += 10
sq = n^2
while sq > powerOf10
# the square has more digits than the previous one
powerOf10 *= 10
addOnes *= 10
addOnes += 1
end
addUnitSq = sq + addOnes
if addUnitSq < powerOf10
# squaring the number with one added to the digits has the
# same number of digits as the original square
local root = isqrt( addUnitSq )
if root^2 == addUnitSq
# the addUnitSquare is actually a square
# make sure there are no 0s in the addUnitSquare as they
# can't be decremented to a positive digit
if all( x -> x != 0, digits( addUnitSq ) )
suCount += 1
suSquares[ suCount ] = addUnitSq
end
end
end
end
println( suSquares )
end
- Output:
[1, 36, 3136, 24336, 5973136, 71526293136, 318723477136]
Lua
Tested with Lua 5.2.4
do --[[ find some sub-unit squares: squares which are still squares when 1
is subtracted from each of their digits
the sub-unit squares must end in 36 (see task discussion) and so
the square formed by subtracting 1 from each digit must end in 25
so, as with the J sample, we start with the "25" square and test
the incremented square
--]]
-- find the sub unit squares - note that 1 is a special case
io.write( "1" )
local suCount = 1
local addOnes = 1
local powerOf10 = 10
local n = -5
while suCount < 7 do
n = n + 10
local sq = n * n
while sq > powerOf10 do
-- the square has more digits than the previous one
powerOf10 = powerOf10 * 10
addOnes = ( addOnes * 10 ) + 1
end
local addUnitSq = sq + addOnes
if addUnitSq < powerOf10
then
-- squaring the number with one added to the digits has the
-- same number of digits as the original square
local root = math.floor( math.sqrt( addUnitSq ) )
if root * root == addUnitSq
then
-- the addUnitSquare is actually a square
-- make sure there are no 0s in the addUnitSquare as they
-- can't be decremented to a positive digit
local noZeros = true
local v = addUnitSq
while v > 0 and noZeros do
noZeros = v % 10 ~= 0
v = math.floor( v / 10 )
end
if noZeros
then
io.write( " ", addUnitSq )
suCount = suCount + 1
end
end
end
end
end
- Output:
1 36 3136 24336 5973136 71526293136 318723477136
MiniScript
ctr = 1
tt = time
while ctr < 1000000
a = ctr ^ 2
aa = str(a)
if a == 1 or (a % 100 == 36 and aa.indexOf("0") == null) then
b = a - ("1" * aa.len).val
c = b ^ 0.5
if floor(c) == c then print a
end if
ctr +=1
while not (ctr % 10 == 4 or ctr % 10 == 6)
ctr += 1
end while
end while
- Output:
1 36 3136 24336 118336 126736 5973136 71526293136 113531259136 137756776336 318723477136
Nim
import std/[algorithm, math, strutils]
func digits(n: Positive): seq[int] =
## Return the sequence of digits of "n".
var n = n.Natural
while n != 0:
result.add n mod 10
n = n div 10
result.reverse()
func toInt(digits: seq[int]): int =
## Convert a sequence of digits to an integer.
for d in digits:
result = 10 * result + d
func isSquare(n: int): bool =
## Return true if "n" is square.
let r = sqrt(n.toFloat).int
result = r * r == n
echo "First eight sub-unit squares:"
echo 1
var n = 0
var count = 1
while count < 8:
inc n, 5
block Check:
var digs = digits(n * n)
for d in digs.mitems:
if d == 9: break Check
inc d
let s = digs.toInt
if s.isSquare:
inc count
echo s
- Output:
First eight sub-unit squares: 1 36 3136 24336 5973136 71526293136 318723477136 264779654424693136
Perl
Like Raku, but without the laziness.
use v5.36;
sub U ($n) { $n>1 ? (2*$n)**2 : 1 }
sub L ($n) { $n**2 }
sub R ($n) { 1 x $n }
my($l,$u,$c) = (-1);
while (++$u) {
next if U($u) =~ /0/;
my $chars = length U($u);
while ($l++) {
next if U($u) - L($l) > R($chars);
last if U($u) - L($l) < R($chars);
say U($u) and ++$c == 7 and exit
}
}
- Output:
1 36 3136 24336 5973136 71526293136 318723477136
A more efficient approach:
use 5.036;
use bigint try => 'GMP';
use ntheory qw(:all);
sub difference_of_two_squares_solutions ($n) {
my @solutions;
my $limit = sqrtint($n);
foreach my $divisor (divisors($n)) {
last if $divisor > $limit;
my $p = $divisor;
my $q = $n / $divisor;
($p + $q) % 2 == 0 or next;
my $x = ($q + $p) >> 1;
my $y = ($q - $p) >> 1;
unshift @solutions, [$x, $y];
}
return @solutions;
}
my $N = 20; # how many terms to compute
my %seen = (1 => 1);
my $index = 1;
say($index, ': ', 1);
OUTER: for (my $n = 1 ; ; ++$n) {
my $r = (10**$n - 1) / 9;
foreach my $xy (difference_of_two_squares_solutions($r)) {
my $xsqr = $xy->[0]**2;
my @d = todigits($xsqr);
next if $d[0] == 1;
next if !vecall { $_ } @d;
next if !is_square(fromdigits([map { $_ - 1 } @d]));
if (!$seen{$xsqr}++) {
say(++$index, ': ', $xsqr);
last OUTER if ($index >= $N);
}
}
}
- Output:
1: 1 2: 36 3: 3136 4: 24336 5: 5973136 6: 71526293136 7: 318723477136 8: 264779654424693136 9: 24987377153764853136 10: 31872399155963477136 11: 58396845218255516736 12: 517177921565478376336 13: 252815272791521979771662766736 14: 518364744896318875336864648336 15: 554692513628187865132829886736 16: 658424734191428581711475835136 17: 672475429414871757619952152336 18: 694688876763154697414122245136 19: 711197579293752874333735845136 20: 975321699545235187287523246336
Phix
Uses the much faster method from the OEIS page.
with javascript_semantics function sub_unit_squares(integer digits, atom d1s) -- d1s is eg 1111 for digits==4 sequence res = {} for f in factors(d1s,1) do atom g = d1s/f, a = (f+g)/2, b = (f-g)/2 if a*a-b*b=d1s then string a2 = sprintf("%d",a*a) if length(a2)=digits -- and (digits=1 or a2[1]!='1') and not find('0',a2) then res = append(res,a2) end if end if end for res = unique(res) return res end function atom n = 1, t0 = time() sequence res = {} for digits=1 to iff(machine_bits()=32?12:18) do res &= sub_unit_squares(digits,n) n = n*10+1 end for printf(1,"Found %d sub-unit squares in %s:\n%s\n",{length(res),elapsed(time()-t0),join(res,"\n")})
As per the Julia jq entry, while A061844 explicitly excludes squares beginning with 1 apart from 1 itself, this task does not.
It would of course be trivial to exclude them using either and a2[1]!='1'
, or and length(sprintf("%d",b*b))=digits
.
Technically 32 bit will go to 16 digits and 64 bit to 19, but it won't find any more answers, so there's no point.
- Output:
(64 bit - 32bit only shows the first 11, in 0s)
Found 15 sub-unit squares in 7.1s: 1 36 3136 24336 118336 126736 5973136 71526293136 113531259136 137756776336 318723477136 118277763838638336 118394723955598336 126825269186126736 264779654424693136
Should you uncomment the additional test so the output matches A061844, it'll only show 8 (or 7 under 32-bit).
gmp version
You can run this online here.
with javascript_semantics include mpfr.e constant match_A061844 = true function sub_unit_squares(integer digits) mpz d1s= mpz_init(repeat('1',digits)) sequence res = {} mpz {g,a,b,t} = mpz_inits(4) for f in mpz_factors({"mpz",d1s},1,false) do mpz_fdiv_q(g,d1s,f) -- g = d1s/f mpz_add(a,f,g) assert(mpz_fdiv_q_ui(a,a,2)=0) -- a = (f+g)/2 mpz_sub(b,f,g) assert(mpz_fdiv_q_ui(b,b,2)=0) -- b = (f-g)/2 mpz_mul(a,a,a) mpz_mul(b,b,b) mpz_sub(b,a,b) if mpz_cmp(b,d1s)=0 then -- a*a-b*b = d1s string a2 = mpz_get_str(a) -- (a*a by now) if length(a2)=digits and (not match_A061844 or digits=1 or a2[1]!='1') and not find('0',a2) then res = append(res,a2) end if end if end for res = unique(res) return res end function atom t = time() sequence res = {} for digits=1 to 32 do -- (next up would be 40, 38 exceeded my patience) res &= sub_unit_squares(digits) end for printf(1,"Found %d sub-unit squares in %s:\n%s\n",{length(res),elapsed(time()-t),join(res,"\n")})
- Output:
Found 23 sub-unit squares in 0.9s: 1 36 3136 24336 5973136 71526293136 318723477136 264779654424693136 24987377153764853136 31872399155963477136 58396845218255516736 517177921565478376336 252815272791521979771662766736 518364744896318875336864648336 554692513628187865132829886736 658424734191428581711475835136 672475429414871757619952152336 694688876763154697414122245136 711197579293752874333735845136 975321699545235187287523246336 23871973274358556957126877486736 25347159162241162461433882565136 34589996454813135961785697637136
Should you set match_A061844 to false, ie include numbers >1 that begin with 1, it'll find fifty of them.
Cheating slightly, but at least matching/verifying A061844 all bar the last:
--for digits in tagset(32)&{40,42,54,60,64,66,72,84} do -- Found 73 sub-unit squares in 4 minutes and 17s
--for digits in {90} do -- Found 3 sub-unit squares in 4 minutes and 18s
--for digits in {96} do -- Found 1 sub-unit squares in 8 minutes and 30s
--for digits in {180} do -- no dice, eventually blows up mpz_factors()
Python
# sub-unit_squares.py by Xing216
from gmpy2 import is_square
def digits(n: int) -> list[int]:
return [int(d) for d in str(n)]
def get_sub_unit(digits_n: list[int], n: int) -> int:
return int(''.join([str(d-1) for d in digits_n]))
def is_sub_unit_square(n: int) -> bool:
if not is_square(n): return False
digits_n = digits(n)
if 0 in digits_n: return False
sub_unit = get_sub_unit(digits_n, n)
if len(str(sub_unit)) != len(str(n)): return False
return is_square(sub_unit)
res = []
i = 1
while len(res) < 5:
if is_sub_unit_square(i): res.append(i)
i += 1
print(res)
- Output:
[1, 36, 3136, 24336, 5973136]
Raku
First seven take about 5 seconds with this implementation. The eighth would take several hours at least.
my @upper = 1,|(1 .. *).map((* × 2)²);
my @lower = (^∞).map: *²;
my \R = [\+] 0, 1, 10, 100 … *;
my $l = 0;
.say for (gather {
(^∞).map: -> $u {
next if @upper[$u].contains: 0;
my $chars = @upper[$u].chars;
loop {
$l++ and next if @upper[$u] - @lower[$l] > R[$chars];
take @upper[$u] if @upper[$u] - @lower[$l] == R[$chars];
last;
}
}
})[^7]
- Output:
1 36 3136 24336 5973136 71526293136 318723477136
RPL
Since brute force is not an option for 4-bit computer languages, we use the method described in the OEIS page, based on the resolution of x² - y² = 11..1
The line OVER →STR 1 1 SUB "1" ≠ AND
can be removed to obtain the OEIS sequence.
FACTS
is defined at Factors of an integer
RPL code | Comment |
---|---|
≪ IFERR FP NOT THEN DROP 0 END ≫ ‘INT?’ STO ≪ ALOG 1 - 9 / DUP FACTS 1 OVER SIZE 2 / CEIL SUB → ones factors ≪ { } 1 factors SIZE FOR j factors j GET ones OVER / + 2 / SQ IF DUP XPON ones XPON == OVER →STR "0" POS NOT AND OVER →STR 1 1 SUB "1" ≠ AND OVER ones - √ INT? AND THEN + ELSE DROP END NEXT ≫ ≫ ‘SUBUNITS’ STO |
INT? ( x → boolean ) SUBUNITS ( n → { SUS } ) ones = (10^n-1)/9 factors = half of the full FACTS list is enough for each factor x = ((factor+factor/ones)/2)^2 if x has same number of digits as ones and x does not contain any zero and x does not start with 1 and x's sub-unit square is an integer then append x to list next return list |
≪ 2 { 1 } WHILE DUP SIZE 5 < REPEAT OVER SUBUNITS + SWAP 1 + SWAP END SWAP DROP ≫ EVAL
- Output:
1: { 1 36 3136 24336 5973136 }
Returned in 41 seconds on a basic HP-48SX... but finding the following SUS (71526293136) took one hour and a half, essentially to factor 11111111111 = 21649 x 513239
Rust
fn is_square_number( number : u64 ) -> bool {
let root : f64 = (number as f64).sqrt( ).floor( ) ;
(root as u64) * (root as u64) == number
}
fn is_subunit_square( number : u64 ) -> bool {
let numberstring = number.to_string( ) ;
let numstring : &str = numberstring.as_str( ) ;
if numstring.contains( '0' ) {
return false ;
}
else {
if number < 10 {
is_square_number( number ) && is_square_number( number - 1 )
}
else {
let mut digits : Vec<u64> = Vec::new( ) ;
let mut num : u64 = number ;
while num != 0 {
digits.push( num % 10 ) ;
num /= 10 ;
}
for n in digits.iter_mut( ) {
*n -= 1 ;
}
let mut sum : u64 = 0 ;
let mut factor : u64 = 1 ;
for d in digits {
sum += d * factor ;
factor *= 10 ;
}
is_square_number( sum ) && is_square_number( number )
}
}
}
fn main() {
let mut solution : Vec<u64> = Vec::new( ) ;
let mut current : u64 = 1 ;
while solution.len( ) != 7 {
if is_subunit_square( current ) {
solution.push( current ) ;
}
current += 1 ;
}
println!("{:?}" , solution ) ;
}
- Output:
[1, 36, 3136, 24336, 118336, 126736, 5973136]
Sidef
Using the method described in the OEIS entry, representing (10^n - 1)/9 as a difference of squares x^2 - y^2 and checking x^2 to see if it satisfies the conditions.
var N = 20 # how many terms to compute
var arr = Set(1)
for n in (1..Inf) {
var r = (10**n - 1)/9
arr << r.diff_of_squares.map{.head}.map{.sqr.digits}.grep {|d|
(d[-1] != 1) && d.none{.is_zero} && d.map{.dec}.digits2num.is_square
}.map{.digits2num}...
break if (arr.len >= N)
}
arr.sort.first(N).each_kv {|k,n|
say "#{'%2d' % k+1}: #{n}"
}
- Output:
1: 1 2: 36 3: 3136 4: 24336 5: 5973136 6: 71526293136 7: 318723477136 8: 264779654424693136 9: 24987377153764853136 10: 31872399155963477136 11: 58396845218255516736 12: 517177921565478376336 13: 252815272791521979771662766736 14: 518364744896318875336864648336 15: 554692513628187865132829886736 16: 658424734191428581711475835136 17: 672475429414871757619952152336 18: 694688876763154697414122245136 19: 711197579293752874333735845136 20: 975321699545235187287523246336
Wren
Apart from the number '1' itself, it looks like the first digit can't be '1' either. In other words, the number of digits in the transformed number must remain unchanged.
import "./math" for Int
System.print("The first 7 sub-unit squares are:")
System.print(1)
var i = 2
var count = 1
while (count < 7) {
var sq = i * i
var digits = Int.digits(sq)
if (digits[0] != 1 && !digits.contains(0)) {
var sum = digits[0] - 1
for (i in 1...digits.count) sum = sum * 10 + digits[i] - 1
if (Int.isSquare(sum)) {
System.print(sq)
count = count + 1
}
}
i = i + 1
}
- Output:
The first 7 sub-unit squares are: 1 36 3136 24336 5973136 71526293136 318723477136
XPL0
func HasZero(N); \Return 'true' if N contains a zero digit
int N;
[repeat N:= N/10;
if rem(0) = 0 then return true;
until N=0;
return false;
];
func DecDigits(N); \Decrement each digit of N
int N, M, T;
[M:= 1_111_111_111;
repeat T:= N-M;
M:= M/10;
until T >= 0;
return T;
];
int N, C, N2, M, M2;
[N:= 1; C:= 0;
repeat N2:= N*N;
if not HasZero(N2) then
[M2:= DecDigits(N2);
M:= sqrt(M2);
if M*M = M2 then
[IntOut(0, N2);
ChOut(0, ^ );
C:= C+1;
];
];
N:= N+1;
until C >=7;
]
- Output:
1 36 3136 24336 118336 126736 5973136